/u

I :_-r:Jq A

1"
l'
I
',:,::
:,
I
I
I
I
I
I
I
I
I
'10 ;
r29A K:IVIL ENGINEERING STUDIES
234 \
i; 3 STRUCTURAL RESEARCH SERIES NO. 234
L AD DEF R 11 N CHARACTERISTICS
F BE maC LU N C NNECTI NS
IN REINF ReED C NCRETE
By
Matz Reference
Civil Engineering Department
BI06 C. E. Buildlng
University of Illinois
Urban&, 61801
N. H. BURNS
C. P. SIESS
A REPORT ON A RESEARCH PROJECT
Sponsored by
THE PORTLAND CEMENT ASSOCIATION
UNIVERSITY OF ILLINOIS
URBANA, ILLINOIS
JANUARY 1962
LOAD-DEFORMATION CHARACTERISTICS
OF BEAlvi-COLUMN CONNECTIONS
IN REINFORCED CONCRETE
by
N. H. Burns
C. P. Siess
University of Illinois
Urbana} Illinois
January 1962
TABLE OF CONTENTS
I. INTRODUCTION . .
1. Object.
2. Scope
3
Aclmowledgment.
4. Notation.
II. EXPER:rnENTAL DATA AND BEHAVIOR .
5. Description of Test Programs.
6. Test Apparatus and Measurements Taken .
7. Behavior of Test Beams .....
8. Modes of Failure at Ultimate.
~ ~ ~
III. , ~ Y S I S .............. .
IV.
V.
VI.
9 . Introduction. . . . . . . . .. ..... ....
10. Moment-Curvature Relationship and Curvature Distrioution.
11. ,Analysis at Cracking Stage. . . . . . .. .
12. Analysis at Yield Stage . . . . . .. ....... .
13 . ,Analys is at Crushing -Stage. . . . . . . . . . . .
14. Analysis at Ultimate Stage . . . . . .
COMPARISON OF COMPUTED AND MEASURED RESULTS. .
1;5. General Remarks. . . . . . . . . . .
16. Tests at the University of Illinois .
17. Ernst's Tests
18. Summary. . .
EFFECT OF VARIABLES.
19. General Remarks . .
20. Effect of Depth.
21. .Effect of Compression Steel .
22. ,Effect of Transverse Reinforcement. .
23 .. Effect of Repeated and Reversed Loading
SUMMARY AND CONCLUSIONS.
24. Object and Scope. .
25. Behavior .....
26. Analysis and Results ..
27. Effect of Variables.
28. Conclusions
REFERENCES . .
TABLES ..
FIGURES ..
-iii-
1
1
2
4
4
8
8
10
12
15
19
19
19
26
29
37
48
55
55
55
60
67
69
69
70
75
81
89
100
100
101
102
109
111
114
115
140
-iv-
TABLE OF CONTENTS (Continued)
APPENDIX A - .MATERIALS" FABRICATION" .AND TESTING . DETAILS
29. Materials ...... , ...... QQ ,
30. Fabrication and Curing Procedure u Q ,
31. Details of Test Apparatus and Instrumentation .
32. Description of Test Procedure . , .. , , . , ,
APPENDIXB -DESCRIPTIONOF COMPUTER PROGRAMS ..
33 u Purpose of 'Programs . . . .. . . . . . , .
34, Input Data and Preset .Quantities, .
35 . Flow Diagrams . . . . , . . . , . n ,
36. Output and '. Estirna tion of Running Time . .
370 Availabili ty. . . . . . QQ.
A.PPENDIX C - LOAD-DEFLECTION CURVES.
206
206
208
210
214
226
226
226
227
231
232
237
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
LIST OF TABLES
Properties of Beams from the Current Test Program. . 0
Properties of Beams from McCollister's Test Program. .
Properties of Beams from Ernst is Test Program .. ' .
Range of Variables Studied inthe.TestPrograms. . .
Computed Kand x' Values for Beams of the Current Test
c
Program. . , . . . . " . . . , . '-, . ... -, . . . .-
Computed K and x' Values for Beams of McCollister's Test
c
Program ..... , , ... ' ....... .
Equivalent Spread Length for Crushing Deflection Analysis by
Method- 3 - Beams of the Current Test Program . 0 , .'.
Equivalent Spread Length for Crushing Deflection ,Analysis by .
Method-3 - Bearris of McCollister's Test Program . 0-' 0 .-
Computed K Values for Ultimate Deflection Analysis -Beams
u , ,-
of the Curren-c Test Program. . . '0 .'. ,; 0
Computed K Values for Ultimate Deflection Analysis -Beams
of McColliter' sand. Ernst's Test, Programs . . . 0 ,'.
Comparison of Yield Moments and Deflections for Beams of the
Current Test Program . , 0 , 0 , .'
Comparison of Yield Moments and Deflections for Beams of
McCollister's Test Program 0 0 0
Comparison. of Computed and Measured Crushing Moments - -Beams
of the Current Test Program . . . . . . , . . . . . . 0 0
Comparison of Computed and Measured Crushing Moments - Beams
of McCollister's Test ,Program ....... 0 0
Crushing Deflections Computed by Method 1 Compared with
Measured Values -Beams o ~ the Current Test Program. . 0
Crushing Deflections Computed by Method 1 Compared with
Measured Values -Beams of McCollister's Test Program ..
Crushing Deflections Computed by Method 2 Compared with
Measured Values -Beams of the Current Test Program. . .
Crushing Deflections Computed by Method 2 Compared with
Measured Values -Beams of McCollister's Test Program ..
':".v-
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
.131
132
-vi--' '
LIST OF TABLES
Number
19
Summary of Crushing Deflections by Methods 1) 2) and
3
for
,Beams of Current Test Program.
133
20 Summary of Crushing Deflections by Methods 1) 2) and
3
for
Beams of McCollister's Test Program. 134
21 Comparison of Computed and Measured Ultimate ,Moments -Beams
of the Current Test Program and McCollister's Test Program. 135
22 Comparison of Computed and Measured Ultimate Deflection for
Beams of the Current Test Program and McCollister's Test Program 136
,23 SummarY,of Computed and Measured Moments at Yield" Crushing"
and Ultimate for Beams of 'Ernst's Test Program. . 137
84 Summary" of Computed and Measured Deflections at Yield,
Crushing, and Ultimate for Beams of 'Ernst fS Test Program . 138
25 Summary of Moments and Deflections for Beams Subjected to
Reversed Loading ,in the Current Test Programs. .... 139
,AI Properties of Reinforcing Bars 216
A2 Summary of Concrete Properties . 218
A3
Summary of Strain Measurements Taken .
219
\. -
Number
1
2
3
4
LIST OF FIGURES
Test Specimens Used in the Different Test Programs
Test Set-up for McCollister's Program ..
Test Set ... iip for Ernst 1 s Program. . 0
Phatograph .of Test Set-up forthe Current Pragram.
5 Generalized Load-Deflection Curve for an Under-Reinforced
6
7
8
9
10
11
12
13
14
Concrete Beam. . .
Typical Continuaus Laad-Deflection Recard Taken with A u t a ~ r a f
'X-YPlotter (Beam J -11). '. . . 0 0 0 0
Typical Deflection Versus Concrete Strain Curves Used in
Determining Measured Crushing Deflectian . . . . .
Phatographs ShawingVisual Appearance of First Crushing.
Phatagraphs ShawingExte;nt of Crushing 'at Ultimate ,for Beam J-10
Without Campression. Steel. . . . . . . .0 .
Phatographs Showing Extent of Crushing at Ultimate for
Beams wi thCampression Steel (A = A') . . . . . .
s s
Phatographs Showing Buckling .of Campressian Steel after
. Failure far Beams with A t = 0.56 A ......... .
s s
Photographs .of Beams 3-6 'and J-22 ShawingShearDisplacement
. ,Associated with Made .of Failure. . 0 0 .; .'
Phatographs .of Beams J-13 and J-20 Showing Shear Displacement
Assaciated with Made of Faiiure. 0 0
Moment Versus Curvature Relatianship for Under-Reinforced
CancreteSection . . . . . . . .. . 0 0
15 Theoretical Distributian .of Mament and Curvature at
17
18
Cracking) Yield) and Crushing. 0
Camparison .of Measured Distributian :af Curvature -at,Yield for
Beam J -17 with Distributian Assumed .in Analysis far Yield
'Defle'ction ;- . , .' . .'. . . . . . ... " ........ ' .. " ....... ... .................. .
Comparison. of Measured Distributian of Curvature at Crushing
for Beam J -13 with Distributions Assumed in Analyses . . . .
Comparison of Measured Distribution of Curvature 'at Crushing
.forBeam J-14 with Distributian.Assumed in.Analyses .....
-vii-
140
141
142
144
145
146
147
148
149
150
151
152
153
154
157
Number
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
-viii -
LIST OF FIGURES (Continued)
Comparison of Measured Distribution of Curvature at Crushing
'for Beam J -17 with Distributions Assumed in Analyses
Idealized Moment Versus Curvature Diagram. . .
Distribution of Moment and ,Curvature Following
M- .cp Relationship . . . .
Distribution of Curvature Along Constant Moment Span .
Effect of Loss of Bond, on Curvature Distribution . . . .
Assumed Distribution of Curvature for Analysis of Cracking
and Yield Deflections ................. .
Relationship Modulus of Rupture and Compressive
Strength df Concrete . . . . . . . . . . . .
Distribution of Curvature at Yield for Analysis of 'Beams
from Ernst1s Test Program ..
Complete Generalized Stress-Strain Curve for Intermediate
Grade Steel Used in Analysis . . . . . . . .
Equivalent Distribution of Curvature forAnalysis of
Crushing Deflection by Method 1 . . . . . . . .
Shape Factor) K) :forMethod 1 Crushing" Deflection Analys is. .
. c
Equivalent Distributions of Curvature forAnalysis of
Crushing. Deflection by Methods 2 and 3 . . . .. . .
Equivalent Spread Length for Computation ,of Crushing
Deflections by Method 2. . .......... .
Additional Equivalent Spread Length for Computation of
Crushing Deflections 'by Method 3 . . . . . . . . . . .
Tension Steel Strains JyIeasured from Punch Marks after Failure ..
Plot of Computed K Values from Ultimate Deflection Analysis
Versuspl/p ... aU
Plot ofMeasl,lredVersus Computed Yielq Moments.
,Plot of 'Measured Versus Computed Yield Deflections .
Plot of Measured Versus Computed Crushing:Moments.
158
159
159
,160
160
161
162
164
165
166
168
170
171
172
173
174
Number
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
-ix-
LIST OF FIGURES (Continued)
Plot of Measured Versus Computed Crushing Deflections
-by Method 1 ................... , ..
Plot of Measured Versus Computed Crushing Deflections
by Method 2 ................... , ..
Plot of Measured Versus Computed Crushing Deflections
-by Method 3. . . . . . . . . . . . . . . . , . . . .
Plot of Measured Versus Computed Ultimate Moments.
Plot of-Measured Versus Computed Ultimate Deflections, ..
Load-Deflection Curves for Beams of -Ernstrs Program ..
Effect of Depth for-Beams with prjp
Effect of Depth for Beams with_prjp
Effect of Depth for Beams with pfjp
Effect of Depth -and Compression Steel-
Effect of Compression Steel for Beams
Effect of' Compression Steel for Beams
Effect of Compression Steel for Beams
o
1.0.
for Beams with #2-Stirrups
with d = 10 in.
with d 14 in.
with d. 18 in.
Effect of Variation in Transverse Reinforcement for
-Beams with d = 14 in . .............. .
Effect of Variation in Transverse -Reinforcement for
Beams -with- d = 18 -in. . .... 0
Photographs Showing Crushing Pattern After Failure for
Beams of the Current Programliaving :U "",Stirrups . .
'Load-Deflection Curve for Beam J-17 Showing Change-in Slope of
Repeated Loading Curve Prior to Initial Yield of the Tension
Steel. . . . . . . ...
Effect of Iamage on Stiffness.
Load Versus Deflection for Beam J-3 ..
Load Versus Deflection for Beam J-7.
175
177
178
179
180
181
182
183
184
185
186
188
190
191
192
193
196
Number
59
60
61
62
64
.Al
A2
A3
A4
A5
A6
Bl
B2
B3
';".x-
LIST OF FIGURES (Conrtinued)
BauschingerEffectinS.A.E. 1030 Steel with
First Half Cycle in Tension. . . . . . . . .
Bauschinger Effect in S.A.E. 1030 Steel with
First .Half Cycle :in Compression ....
Load.Versus:Deflection for Beam J-12
Photographs Showing Different Modes Failurefor
Companion Beams with and without Reversed Loading.
Effect of Reversed Load:lngonDuctility as Shown by Comparison
of the Downward Load-Deflection Curves for Beams J-3and J..,8
with d = 10 in. . ......... 0
'Effect of Reversed Loading :on'Ductility as Shown by Comparison
ofthe.Downward Load-Deflection Curves for BeamsJ-6 and J...,7
with d = 18 in. . ...................... .
Effect of'Reversed Loading on Ductility as Shown by Comparison
of the Downward Load-Deflection Curves for Beams J-5 and J-12
with .d = 18 in. . 0 0 0 0 0 0
Typical Str.ess-Strain Curve for' Reinforcing Bars . Used.in the
. Current Test Program 0 0 0
Photographs of'Specimens Used in Compression Test of
. . . '0
Comparison.of (J-E Curves Tension and Compression.
'Details of 'Stirrups Used in Beams of the : Current Test Program. .
Cross Section of Load Dynamometer
,Photographs. Showing Mechanical Strain Gage Plug Locations. 0
Flow Diagram for Yield Stage Computer Program. 0 0
Flow Diagramfor Crushing Stage Computer Program
. Flow Diagram for Ultimate stage Computer Program .
199
200
202
203
204
205
220
221
222
223
224
225
234

236
I. INTRODUCTION
1. Object
The objective of this investigation was to" establish lOad-deformation
relationships for the beam-to-colurnn connections of relnforced concrete frames.
The approach has been mainly experimental) guided by theoretical considerations
and the results of previous investigations.
More -research has been done in connection with strength than
deformation .of reinforced concrete frames. There is now considerable
interest in extendinglimit analysis to reinforced concrete frames. If we are
to determine the applicaqility of this analysis to these structures) the
deformation characteristics as well as the strength of the . members must. be
known. The assumedredistribution of mom,ent in . limit analysis -requires Buf-
ficient deformation capacity to allow it to occur without some . sort of premature
failure.
There is a need informatioD: on strength and,ductility in
connectiC?n the'behavior 'of frames subjected to earthquake or 'blastloading)
. since the energy absor}?tion capacity of amemberof the structure is a function
of the area under curve for that member. The tests reported
. "
herein dealt only with static loading) but the :resultsshou.ldshedlight on the
'necessq,ry to insure strength and ductility of frames
under other . loading .condi tions .
*
Gaston (1) tested 33 members with third-point loading to determine
load-deformation in pure flexure up to the point of failure.
McCollister (2) reported 25 tests on beams which were simply supported and
*
Numbers ,refer to ,entries in the List of References.
1
-2-
,loaded at 'midspan through a column stub. Ernst (3) tested 20
beams which were loaded through a column stub of varying length atm:j..dspanwhile
maintaining a constant spar ,length. These tests encompassed a range of
.variables, but the depth was maintained as a ,nominal 10 in. for all tests. This
made it difficult to draw 9.ny real conclusions regarding the effect of depth,
which was a major variable in the tests reported ,herein. ,
2, ,Scope
In order to study the load-deformation ,characteristics of heam-column
connections, tests were made on beams simply supported all.d loaded through a
column,Btub at midspan," Thecolurnnstub was extended both 'abov'e and below
,the members in these tests , of the specimens -reported by McCollister
. ,
,had 'no stubextending'below the bottom of the member' (Fig. 1).
Testsweremade'on 21 beams having ,the following 'prfrnary varfables:
1. ,Effecti veDepth -,-d = 10, 14, and 18 in .
2.' Compression: Ratio' -- plip = 0, 0.56,' and 1.0.
,3. Loading Be <luence 'and Direction-,- Repeated arid Reversed. loadIng .
The concrete 'strength varied from f'= 3590 psi tofT =5160 -psi; this'was
c c
a .secondary variable ,mostof the beamshavingab'out4500ps i eoncrete. The
same mix wasus'edfor all members but moistu-re' content of aggregate and age of
'beam at testing 'Varied' somewhat. All steel used. in these tests "was inter-
mediate grade and had essentially the same yield 'point, although there was some
variation. Properties of beams from the current test program are given in
" ,
Table 1. Appendix A contains a more' detailed d'escription of the mater'ials
used and test procedures followed. in' thec'urrent program.'
In order to broaden the scope of this study, beams 'reported by
McCollister -and ,Ernst were studied along with those described above.
'beams all had a, nominal depth of 10 in. and a width of6 in. The concrete
--3-
strength) percentage of tension reinforcement, and percentage of compression
reinforcement varied over a wide range in McCollister's tests. Ernst1stests
involved a 'variation in stub length arid rate of 'loading. Also, McCollister1s
and Ernst
1
s beams were tested on a span of 9 ft. while those tested here were
supported on a12-ft. span. Tables 2 and 3 show summaries of beam properties
from the investigations by McCollister and Ernst, respectively.
In order to insure against shear failures, closed #3 stirrups at
6-in. spacing were provided for all but four of the beams of the current
program as shown in Table 1. These stirrups also served as binders ,to confine
a concrete core at ultimate) and the constant size and spacing of stirrups
eliminated this 'as a variable in these tests as much as possible. The sup-
port given to the compression steel is important since it influences the
buckling 'strength at ultimate. The compression steel itself also confines
the concrete J and the cOnstant spacing of stirrups supported this -steel at
constant intervals ,for all tests. McCollisterlstest specimens also were
provided with #3 stirrups at 6-in. spacing except for a ,few cases as shown in
Table 2.
The behavior of a beam with increasing deformation is a constantly
varying process) but a few points along the w-aymark significant changes
which are important. These -stages of behavior are thecri tical points 'which
have been considered in behavior and analysis throughout this study. In the
order of'occurrence they are:
(1) Cracking of the concrete, marking ,the change in stiffness from
that of an uncracked to a cracked. section.
(2) Yielding of the tension reinforcement) with the beginning of
inelastic action--increase in deflection with ,little increase
in load.
(3) First crushing of the concrete) indicative of first compression
damage to the member and representing :a stage at which analysis
of 'load and deformation can be made.
-4-
(4) Ultimate load-carrying capacity, corresponding 'with -extensive
concrete' crushing -and. buckling :of the compression at-eel inmost
cas-es -where it is -present in reasonable percentages . If too
much-compression steel is 'provided,the tension steel'may'
actually fracture before the compression steel buckles. This
point is well, beyond the -first crushing :stagefor- under-
-reinforced -beams} e:specially provided with compression
steel.
In this 'study analysis was made at each of these stages , and acceptable
agreement between computed and measured values was obta;i.nedformoments 'and
deflections. The combination of t-estsin this series -with those previously
reported produced a :ratherwide -range of variables, and the 'results are more
general than mightpe the case with any. one of the studies alone.
3. Acknowledgment
The test-s 'and analysis 'of ;results 'reported herein were a part of an
investigation of s-t;rength and deformation characteristies of reinforced concrete
beam-column connections conducted .inthe Structural Research Laboratory of the
EngineeringExperiment Station of the University of Illinois. This project was
'sponsored by the Portland Cement Association.
The project director wasDr.C. P. Siess and the experimental work
was 'carried out under the supervision of-Dr. Arthur Feldman. This "report was
written asa Ph.D. under ,the direction of D,r. C. P. Siess, 12rofessor
of Civil Engineering.
The author -is deeply indebted to the following 'Research Assistants
in Civil Engineering 'Who ass.isted with the laboratory testing -and data re-
duction for this 'project :-Einar Bkjorten, -H),Garda- Yamashiro, and Peter
Gergely. Appreciation is also due to Dr. M. A. Sozenforhis help:ful suggestions
during the course of this 'study.
4. Notation
Thefollawing -notation is used in this -report:
.. 5-
A area of tension reinforeement
s
Af area of compressionreinforeement
s
b width of rectangular beam
C compressive force in concrete
c
C
1
compressive force in compression steel
c
d depth from compressive face of beam to centroid of tension reinforcement
d
f
depth from c0p1pressive face of beam to centroid of compression
reinforcement
Dr diameter of compression.steel
E modulus of elasticity of concrete; assumed approximately equal to
c
'E
s
f!
C
f
eu
f'
cu
f
s
fY
S
f
Y
f!
Y
f
u
f'.'
u
I
kd
kd
c
k d
u
=.
-
this study where both Ecand inksi
modulus of elasticity of reinforcing -steel assumed to be 30)000 ksi
compressive strength of concrete as determined from tests of
6 x12-in. cylinders
average concrete stress .at crushing stage
average concrete stress at ultimate
stress in tension reinforcement
stress in compression .reinforeement
yield point of tension reinforcement
yield point of compression reinforcement
ultimate tensile strength of tension reinforcement
ultimate tensile strength of compression reinforcement
moment of inertia of gross concrete section
depth to neutral'axis of transformed section (straight line theory)
depth of stress block in concrete at crushing -stage
depth of compression block below eentroid of compression steel at
ultimate. [Or (k d + d! ) = depth from compression .faee of original
u
beam. ]
coefficient defining theposition of the internal compresSive foree
in the concrete
K
c
K
u
L

M
Y
M
c
M
u
11
p
p'
~
T
w-
a
f3
8-
y
6
c
b:.
u
E
s)
E
sy"
E
sh"
Eu"
E
C
E
cy
r!
Co
S
-6-
=shape factor coefficient in Method 1 analysis for'crushingdeflection
shape factor coefficient in analysis 'forultimate deflection
length of beam span
distance from support to face of coluln-Yl stuD
bending moment at yield .stage
bending moment at crushing 'Stage
bending moment at ultimate stage
E IE = modular 'ratio
c c
As/bd tension steel ratio
A'lbd compression steel ratio
s
p f If!
Y c
force in tension .reinforcement
width of column stub
(1 . --M 1M ) . in Method.l analysisforerushing deflection
y 'c
(1 - -M 1M ) in analysis' for ultimate deflection
c u
midspan deflection at yield stage
midspan deflection at crushing :stage
midspan deflection at ultimate stage
strain .in tension and compression steel) respectively
6' strain in tension and compression steel" ,respectively" at yield
.point
sy
6'
sh
strain in tension and eompression steel" respectively" at work
hardening
E'
u
ultimate strain at fracture in t'ension and compression steel }
respectively" from tension tests.
concrete strain at top; fiber at :crushing stage
concrete 'strain at top fiber 'at yield stage
ultimate strain for confined concrete core at level of compression
steel (6'= 6' forbeamswith compression steel)
cu su .
-7-
qJy
= computed curvature df 'beam at yield stage
' ~ c
computed curvature of beam at crushing stage
~ u
computed curvature of 'heam at ultimate stage
II. . EXPERIMEl\4""TAL DATA AND BEHAVIOR
5. Description of Test Programs
Experimental data from tests of 59 beams in three different programs
have been utilized in order to provide as wide a range of variables as p8S-
sible for thisstudy. Tests 'of18 of the beams were carried out at the
University of Illinois during '1952-1954 and were reported by McCollister in
June 1954 (2). The original test data were available from these tests. Twenty
of the beams were from a <series of related tests conducted by Ernst at the
University of Nebraska which were reported in June 1957 (3). The current test
program, involving 21 beams, was conducted by the writer at the University of
Illinois during '1959-1961.
In thissect:ion and the following :section the types of specimens
tested, the test setups used, and the measurements taken in these three test
programs are outlined. Tables 1, 2, and 3 contain the essential details of
all the test s-pecimens. Details of the materials used and test procedures
followed by McCollister and Ernst are :given in the -reports of their
investigations. Appendix A contains these details for the current test
program.
McCollister's beams were all 6 by 12 in. in cross section and 10 ft:
long with a span between supports of 9ft. The load was applied through a
6 by 6by12-in. column stub cast integrally on the top of the beam at midspan
as shown in Fig .. la. The principal variables in this program were : tension
steel ratio, .p,; compression steel ratio, p' ,; ratio pl/p,; and concrete strength,
f". Table 4 gives the range of values .forall of the variables involved in
c
McCollister's study.
-8-
-9"
.stirrups consisted of #3 deformed a.re.ctangular ,loop
with both ends hooked around.a.longitudinal baron the tension side of the
beam. The s:pacing was 4 or.:6 in .. as shown in Table 2 along with other
of the beams .
. Ernst's beams were also 6 by 12 in. in cross and 10ft
long , with a span of.9 .. ft. . Column stubs '9-1/2. in .. high and 6.in. wide were
cast integrally on top and bottom of each at midspan, the length of this
stub being :oneof the variables .. load was applied tJ+rough the colunm
stub as shown in Fig .. lb. The principal variables..in tnisprogram were:
;','i
tenE?ion steel ratio ,p; length .of COJ,unm stub" w; and ,:rate of 'loading. Table 4
gives the range of values' for 'all of the variables . involved in Ernst,' s study.
Stirrups were made of #3 bentin:to rectangular
and spaced'SO that all the 'predictedmaximum spear could.be 'carr-iedby the
stirrups at a unit stress below point. The stirrup spacing:is.shown
in Table 3 along'with other 'properties 'of the beams.
The test specimensused for the current program .were of varying
, '., . .' .'--:---.. .. --' . ..:. .. .. .....
'cross sections 'with depth being -a "primary variable. All but two of the beams
were 8 in. wide 'and '13 ft long 'witha span of '12 ft, thewidthbeing6:ln. in
_,_ ...._._ ... . ........ ".- . , .. ........ _ ....,-."" .." ....... ...........t, __
two cases. .Theload was applied, through a column .stub6in. nigh .and .. 12 in.
long whichwase<1ual in width to the beam cross 'section and cast 'integrally
on top and bottom of each beam at .. midspan (Fig ",I.c) . The. primary variables
in this program were : .depth, d; compress:Lon steel percentage J . p'; ratio
p'/p; and loading :seCluence and direction. Table 4 gives the range of values
for 'all of the variables involved in the current testprogra;m'
ExceJ?tfor four. 'beams in which. #2 plain bars 'were the stirrups
. for the current program consisted of #3 deformed bars bent into rectangular ..
:ties' and welded closed. The Bpac-ingwas6 'in. for -all beams as shown .in
Table 1 along with other 'proJ?erties of the beams.
-10-
6. Te:stApparatus and Measurements Taken
McCollister I S heams were.loaded in a 300} OOO-lb. capaci ty Riehle
screw-type testing machine with the arrangement shown ,in Fig. 2. Load was
measured oymeans of a 50" OOO-lb . capacity 'ring 'dynamometer 'placed immed1at'ely
above the column stub.
Strains in the tension steel were measured on 6-in .. gage lengths
,through core holes us ing'a Berry type mechanical gage. The 'gage lines extended
continuously over the full span of the beam.
Compressive 'Strains in the at 'the top surface'were measured
with Type A-ll SR..,.4electric 'st.raingages having agage.length ofl in. ,The
gages were arranged :.indifferentpatterns ,in the 'vicinity. of the column stub
,face to detect peak strains and.toaid .in detect,ing ::firsterushing.
Compress'i ve steel strainsWeremeasured. wi thType A-ll SR-4electric
. .
strain gages 'mounted' andwaterproofedbeforethe :contretewas cast. These
gages were located at the 'faceofthecolumri 'stub.
, '
The deflectionswi th :respect to the bed of the testing machine were
measured 'at s'everal points along "the span with aste.el peale. A dial indicator
was used to measure the midspan deflection.
,Ernst's beams were ,tested on'a'speeiallY designed .supportin a
400,000-10. hydraul'ic t,estin'g,rhaO.hinewi th fixed ,loading ,head as shown in
Fig. ,3. The testing machine was used to 'measure the load for ,the slow tests
and Baldwin Type C 'load 'cells 'wereusedfor the fast tests .
. Strains in the by. means ofTypePA-3
(post yield) SR-4 electric strain gages having :agc;Lge lengthof '3/4 in. placed
through core holes after'curingwcis eompleted. T.hese 'gages were located at
midspan and at . points just outs,idethe 'eolumn :stubfor ,the slow tests} but
only at midspan ,for the fast "tests.
-11-
Compressive -atTains 'in theconcTete at the top surfaeewere-measured
with Type A-I SR",,4 eleGtricstraingages (13/16 in. gage length). The arrange-
ment varied somewhat);. but they were located .3/4 in. to 2-3/4 in. away from the
column stub face. In.some cases the gages were mounted 3/4 in. out from the
column stub ;faceand 3/4 in . down from the top surface.
Def'lections were measured by means of an engineer I s level s ighted at
scales mounted at each face of the column stubexc-eptfor . the .beams having a .
6-in. long:stub. In these 'cases, ,the scale was located at mids:pan. The eenter
line deflect.ion was measured with a llcolurnn-type transducerll for fast
tests ..
Forthebeams loaded at the :faster Tate, the duration of the test t-o
maximum load was j to 5 minutes. Dual channel osc-illographswere used to
record the.data.from these tests.
The beams 'of the 'current program were frame) the
load being 'applied by means of a . 50-ton . capacity .hydraulic -rams. ,A photograph
of the test frame ,with abeam in place is shown in Fig. 4. The beam was 'simply
supported by end yokes which carried of the
beam by means of eam yoke :rollers carried on hE:at treated steel shafts. End
reaction supports provided milled horizontal surfaces for these -rollers to
bear -against ,both upward and dOW11,ward. The load was measured by means of
'dynamometers designed and built for this investigation,. as described ,in
Appendix A.
All of, the ,bearnsof the current program were subjected to unloading,
and reloading :at several points duringthe test, and Some of the ,beams were
subjected to reversal of loadingin addition tothe loading.
Strains in the tension steel were measured with TypeA7-4 (1/4' in.
gage length) or TypePA-.3 (3/4 in. gage length - Fost yield)
-,12-
strain gages, 'mounted on the ,bars 'and waterproofed before the beam was cast.
These gages were'located at seve"ral'points in the vicinity of the column stub J
with gages one inchou.tfrom the faeeof "the ,stub in every' case. '
Compressive strains in the concrete ',at the top surface were measured
with Type A ~ 3 SR-4 electric s:traingages having-a gage length "of 13/16 in', The
first gage was, always one inch out from the face of the column stub'andmidway
between.the"vertical faces. Several arrangements of additional gages in the
region adjaeent to the -stub ,were used to rrieas-urethe"profile of concrete
strain at "the top surface 'of' the beam atiricreasing,loadlevels.
Compression steel strains "were measured with Type A 7-4 or Type- PA-3
SR-,4 'elect'ric 'strain gages mounted ,on the bars and waterproofed before the
beam was cast. Thes',egag:eswerealso locat,ed one ,inch out from the column
stub face in every case) ,some beams havi:p.g -additional gages at Other' 'positions
farther' out. from the column 'stub ~
" Deflections were measured at four 'points along ,the -spanW;i th dial
gl3-ge indioat,ors c-arried by a, deflectionhridge supported at the end reactions ~
The 'dials acted,'ragainst -short pieces of angle bounded to the face of the beam
at'mid...,height.
lri' addition ,too the load and' deflectionmeasu.remehts taken at
individual load points during ,the test) continuous.load-deflection curves
were plotted with an Auto'graph Model 3' x-y Plotter throughout' the test.' ' The
load measured with a dynamometer was the llyll input to the plotter) and, the
midspan def'lectionmeasured with a heltcal slide wire displacement transducer
was the llXll input to the plotter.
7. Behavior 'of 'Test Beams
Since the beams considered in this s t u ~ y were all under-reinforced
in the sense that the tension steel yielded before the concrete crushed, the
-13-
significant points in their behavior followed a ,typical sequence. A
gene,ralized load--deflection. curve for an under-reinforced concrete beam is
shown in Fig. 5. Points 1 J 2" 3" and 4 on this curve show the four behavior
stages which define the load-deflection response of the beam; and these same
stages will be considered .in computing :the moments, curvatures, and distribu-
tions of curvature in tqe analys.is of Chapter III . In the
the order of occurrence of these points in the load-deflection history of a
beam will be followed.
At the beginning of the test, the beam was uncracked and the initial
slope of the load-deflection curve was related to the stiffneps of the. gross
section or,proba1;>ly,moreexactly, to the stiffness of the transformed section.
The first break in the curve occurred when the beam cracked in the tension
zone. Beyond this point, more or 'less Gontinuously from
that of the uncracked section to that of a cracked section, and the slope of
the. load-deflection curve changed. accordingly. Point 1 on the generalized
,load-deflection curve in Fig. 5 represents this stage of fir$t cracking.:in the
beam behavior. A small shrinkage crack was usually p;resent at each corner
of the stub before load was applied to the beam,. and the first tension crack
was usually a continuation of this crack at the critical-section. Thisfirst
flexural crack extended well into the beam in a very short interval of deflec-
tion so that the hrea,k.in the load deflection curve was rather sharp. The
continuous load-deflection curves plotted in the current test program showed
this break very clearly (see Fig. 6).
Thesecond Significant stage in the load-deflection,history was the
yield point for the member (Point 2 on Fig. 5).. Since the .members were under-
reinforced, the tension steel the crushing str.ain was reached
.in the concrete. The yield point .was of special importance since it marked

I
J
-13-
significant points in their behavior followed a :typical seSLuence. A
gene,ralized load .... deflection. curve for an under-reinforced concrete beam is
shown in.Fig. 5. Points I, 2, 3, and 4 on this curve show the four behavior
stages which define the load-deflection response of and these same
stages will be considered in computing :themoments, curvatures,and distribu-
tions of curvature .in tqe analysis of OhapterIII. In the
the order of occurrence of these points in "the load-deflection history of a
beam will be followed.
At the beginning of the test, the beam was uncracked and the initial
slope of the load-deflection curve was related to th.e stiffness of the gross
section or ,more exactly , to the stiffness of the transformed section.
The first break in the curve occurred when the beam cracked ,in the tension
zone. Beyond, this point , the 'reduced more or 'less continuously from
that of theuncracked section to that of a cracked section, and the slope of
the load-deflection curve changeFLaccordingly. Point 1 on the generalized
load-deflection curve in Fig . 5 represents this stage of first cracking ': in the
beam behavior. A small shrinkq,ge crackwasus1J.ally present at each corner
of the stub before load was applied to the beam,. and the ,first tension crack
was usually a continuation of thj.scrack at the critical-section. This first
flexural crack extended well into the beam in a very short).nterval of deflec-
tion so that the break in the ,load deflection curve was .rather sharp. The
continuous load-deflection curves plotted.in the current test program showed
this break very clearly (see Fig. 6).
'The 'second significant stage in the load-deflection: history was the
yield point for the .member (Point 2 on Fig. 5). Since ,the members were under-
reinforced, the tension steel yielcied before the crushing str,ain was reached
in the concrete. Theyi,eldpoint .was of special importance since it marked
1
I
)
the boundary between elast'ic -and, inelasticbehaviora.s 'observe'd from the
'beam:' With intermediate grade' tensionre-
inforcement .there was very little increase in load asdeflectionsiricreased
beyond the yield point until steel strains were into the strain hardening
range and deflections were s'everai times the yield value. The yield point'
was eas'ilY observed from thecontinuou's plot of 'load taken in
the curre'nt program (Fig .-6) .
Point 3 on the . curve of 'Fig. 5 represents firstcrushingof the'
concrete. 'This point ,in the load-deflect.fon 'hi'story was not so well defined
in the behavior of the beam as the cracking and' yield points preeeding:it.
- .. .
That is, the load-deflection curve be usedtb detect'this point
sinee there was no significant change in the 'curve at this point. In fact,
the visual appearance of crushing was a gradual one so that it was difficulf
to be very precise in calling a certain deflection the value measured at, first
crushing. Beeause of this, the concrete strain ,gage -readings were also used
tointerpretths .point. . 'rhe gages 'mounted on the top' face of the beam one
inch'outfrom the stub"'face ,shoWed increasing c'ompress,ive strains with
increasing :deflec'tions until crUshing .occurred. The readings reversed at this
point and the plot'of deflection versus strain showed this reversal
rather 'clearly. The curves shown ,in Fig. 7 are typical of these deflection
versus concrete strain curves. 'Inmost cases the first crushing occurred at
about the 'same time on both -sides of the stub J with one 'side leadirigslightly.
This sequence would be expected,especiallysinceloading:througha pin did'
not f'orcethe stub to remain vertical. If yielding of the tension steel
. progressed ,faster onone side thanthe'other, the crushing occurred first on
that side, followed by cr"Lishing' 'onthe6thers ide as deflections increased.
The first' crushing stage . was Vis,ilai stage thana' change in
'load carrying behavior. The damage was in the .form of surface spalling on
-15-
the top of the beam adjacent to the stub. Figure 8 shQws photographs of the
appearance of typical first crushing. This spalling 'usually was noted first
atthe corners of the 'E?tub, gradually increasing 'in its extent as deflections
increased) until the strain reversal on the concrete gage indicated crushing
at the center 'of the top face.
Point 4 on the generalized load-deflection curve of Fig. 5 is the
ultimate load capacit.y for the beam. There was a gradual increase in load
with increased deflections beyond the first crushing stage due to strain
hardening :in the tension steel. Crushing became :nloreextensive, and extended
to the top of the stirrups and compression steel before the maximui1l1oad WdS
reached. The photographs 'in Figs. 9 and ,10 show ,the appearance at this stage.
Inelastic behavior 'Of the concrete produced lateral deformations at 'this
stage. This deformation tended to cause 'Strains in the closed stirrups and
the compression reinforcement, and to this action there was a corresponding
reaction which confined the concrete core within the stirrups in a manner
similar ,to the behavior of a spiral column. The modes of failure differed
according to the amount of compression steel provided. The following section
deals with these modes of
8. .Modes of 'Failure at Ultimate
Reference 2 gives a discussion of the ,modes of failure observed,in
the beams of 'McCollister! sprogram, and the same behavior was observed for
,the beams of the current test program which had similar cross-sectional and
material properties. However, some of the beams of the current program
exhibited somewhat different modes of failure. In this section the ,discussion
of modes of failure at ultimate is concerned with the observed behavior of the
beams of the current program. Since beams having different amounts of com-
pression steel were included here the different modes of failure observed
:-16-
represent the ;range of 'behavior which might .be expected' at"ultimatefor
under-reinforced beams ..
Forbeamswithout compressive reinforcement the internal compressive
force was carried by the confined concrete 'core after crushing had become very
extensive. Figure 9 shows the extent of crushing at ultimate ,for beam J-10
which had no compression steel. . When this core could no longer carry the com...,
'pressiverorce) failure resulted with a rather -rapid decrease in load carrying
capacity as deflection incr.eased. The. crushing of the 'concrete between stirrups
showed the confining effect of the welded closed stirrups closer spac,ing' of
the closed stirrups .in the region of greatest.curvature would result in greater
deformation capacity before reaching the ultimate load due to the increased
confinement offeredthe concrete core.
The use pf compression steel 'increased the ductility of thebeatns
as compared to the companionbeam.s without: compression steel. ' This effect is
discussed .in o_etail in Section21; The internal compressive force was carried
bybotllthe concrete and the . steel when compress.ion reinforcement was present.
As crushing 'of the concreteprogresse.d:,thereinforcement was forced to pick
up.alargerproportion of the total compressive force until the steel reached
the yield point. At ultimate) the compression steel strain was 'well into the
yield range) and. only the confined core of concrete was left to act with the
steel in carry-iug the compression force. In this case" the 'compression steel
also acteo_ with the stirrups in confining t,he concrete core as shown in the
photographs of Fig. 10.
The mode of failure at ultLrnate involved buckling of the compress:ipn
steel. the beams with pllp = 0.56 (#6 top bars and #8 bottom
#6. bars buckled within a 6-in. stirrup spacing as shown in the photographs of
'Fig. 11 for beams J -2 and J -14. The buckling of the steel transferred more
-17-
load to the concrete core than.it could carry) resulting :in cTushingof the
concreteand.a rapid decrease in load as buckling continued. ,The concrete
cover was gone by this stage 'So that the steel was not restrained from buckling
outward as shown in the photographs. Lateral deformation of the concrete core
tended to force the bars outward and added to the likelihood of this mode of
failure.
For .'beams having two #8 bars top and bottom) the mode of failure at
ultimate was slightly different. The larger -steel area carried a, greater
proportion of the total compressive force and at the sametimeprovided .greater
confinement for -the concrete core. In general, both the deformation capacity
and the ultimate load were increased by these factors. ,Failure occurred when
the concrete crushed and the 'steel buckled) but this buckling was accompanied
by a shea;ring:movementalong an inclined crack for the deeper beams as shown
in Figs.:12and, '13.FigureI2a shows a photograph of the buckling configura-
tio:q. ,the compression steel of 'beam J...,6. The large shear de:fbrmation in
connection with this buckling mode may be noted from the movement of the
original beam
It was hard to determine the exact mechanism of failure in some
cases. The concrete ,<;:!ore showed extensive crushing and the steel had buckled)
but they appeared to occur 'almost simultaneously at ultimate. Grushingof
the concrete probably was the ,initial cause of failure followed by buckling
of the compression steel. The shearing movement was a part of the failure
mode of the deeper beams 'w;L th #8 top steel) but this probably resulted from
the loss in shear strength due to excessive concrete crushing. The stirrup
was forced to carry the shear; and since it was unable to restrain this shearing
motion) failure resulted. The preliminary investigation using #2 stirrups
rather than #3 stirrups confirms this shear-buckling mechanism which is dis-
cussed.in more detail in Section 22.
--18...;
_One beam in this investigation, J-5, failed by fracturing-the tension
steel, and several of: those tested by McCollister failed _in this manner. For
ratios of pl/p greater than 1.0 and reasonable stirrup spacing and concrete
strength, this Ivould be a very likely mode of failure. If the compression
st-eel area is considerably greater than -the tension steel area, it would be
possible to fracture the tension steel before the compression steel was
stressed to its -yield point .
III. ANALYSIS
9. Introduction
The behavior of the beams tested in this investigation has been
discussed in Section 7 in terms of four significant points which define
the load-deflection .response of the member. In this section a method of
analysis for 'momentand deflection of an under-reinforced concrete member at
each of these four stages 'willbe presented. In the analysis it is more con-
venient to deal with moment rather than with load at each of these stages of
behavior; once the momentsarekriown at the critical section of the simply
supported test .specimenused ,in these tests, the .loadsare easily determined
from static'S. The deflection at each stage is a function of the magnitude
and distribution of curvature along the span.
Section 19 deals with the theoretical relationship between moment
and. curvatureforanunder-reinforced concrete -section; The distribution of
curvature along the span according :to the usual theory at the various stages
is compared with the actual curvature as 'measured in tests of beams in the

.Inthe analysis ateEl,ch of the four stages presented in Sections 11
through'14, the methods of computing'moment and curvature at the critical
sect'ion as well. as the assumed distribution of curvature along the span used
to compute the deflection aregtven. A comparison.of thecomputed and the
measured values for 'moment and deflection using :thisanalys.is is presented
and discussed in Chapter IV.
10. :Moment-Curvature Relationship and OurvatureDistribution
The bending moment diagram fora simply supported beam is known
from statics. If the bendingmoment versus curvature relationship is known
-19-
-.20-
for the beam c-rosssectionateverypointalongthe span) !then the 'load-
deflection response of the'membercanbe detennined .. Fora reinf0rced c-oncrete
member having and material properties along theentire
length, this moment versus curvature relationship for the cross section may be
determined if we know 0r can assume the following:
.1) Dist:r'ibution of strains over thedepthof the cross section
2) Stress-strain relationsJ:?;ipfor ,the, concrete
j) St-ress-strain :relationship ,for 'the steel
4) Equilibrium of theoross 'section.
The distributi0n0fa.trains over thedepth0f the section has been
assumed to be linear of reinforced c0ncretesectibns.
Experimental measurements have shown this to 'be a reasonable assumption when
I
strains 'aretakenover gage.lengthssuffi'cientlY,long to average the strains.
The' stress-strain ';relationships .for 'concrete and. steel, are -.rather well !mown
concrete and couponte.nsion tests on steel) and the
materialin the. member. is assumed to behave the sameas.the test samples.
Equilibrium of the cross section is simply a matter of satisfying the support
condition of the simple beam) but this condition is essential to the analysis.
An -under-reinforced concrete se c.t ion has a moment versus curvature
relationship as shown 'qua;Litatively in Fig:.' 14. 'rhe first break in the .curve
co-rresportds to oracking :of the concrete) and.' the secqndb;reak to yielding 'of
the tension steel. The curve is, shown toorushing :of theconerete
;
in this diagram althoughacons:iderable range beyond crushing :actually exists.
The beam of . Fig .. 15 is assumed to haveE\t each cross section the
-
m0i:nent-curvature relat:iob;ship shown ,in .Fig. ,14. As: the load is increased .. the
b;endingmoment at the c:ri tical section folloW's the se<;luenceshoWn:
cracking).yield arid .crushing. Thebendirtgmomentdiagram cGrresponding to each
-21-
level of load.is statically determined. Initially the beam is uncracked, and
the behavior 'is essentially elastic as the load is . the-level pro-
J
ducing the cracking moment, M J at the critical section. At this stage the
cr
distribution of curvature is as shown in Fig. 15a
J
the curvature at eracking
being known from .themoment-curvature relationship of 'Fig .14.
After cracking, the -slope of the moment-curvature diagram is flatter
than the .initial portion of the curve. When the load :isincreased to a level
sufficienttoprodlice the yiel.d moment J My) at the oritical section, the dis=
tribution of curvature is asshown.in,Fig. 15b. In this case the distribution
is not linear butfollows the moment-curvature relationship of Fig. 14 with
the occurring at the point in the span where the cracking moment, Mcr'
is present.
When thelo9.d is large enough to produce the crushingmo;ment, M ,
c
at the critical section) .the curvature distribution ,is as shown in Fig. 15c.
This distribution -moment-'curvature relationship of . Fig. 14)
withhreaks occurring'atthe two points in the span corresponding to M and
cr
M from the moment diagram.
y
,The curvature distributions fro:r;n the analysis above maybe
compared with measured distributions at the different stages asa check on
the validity. of the analysis. The distribution atfirsteracking shown in
Fig. 15a -.is a good approximation to the fbrm of the measured distribution at
deflections slightly. larger than ,first cracking .. At crackingthere arerrtany
uncertainties .in theanalys is which are more 'S'erious than the error . in this
assumption for curvature distributioJ:]..Section 11 deals with analysis at
cracking stage .
. The measured distribution of curvature at follO\,fs the form
predicted by the analysis except in the immediate 'vicinity of the critical
-22-
section. Figure 16 shows 'the measured distribution of curvature at yield
for one Gf the beams of this ,1nvest,igation.This distribution is very similar
to the theoretical curvature distribution at this stage (Fig .. 15b). The break
in the theoretical distribution at the point where the cracking'momentis
present .is .lostin the accuracy of measurement of the real curvatures, but the
error which results from neglecting this 'point and assuming :a linear -distribu-
tion along the 'Span .isnot significant. The measured peak curvature in the
immediate vicinity of the c-ritical section results from localized yielding ,in
the tensionsteel at this point before the load"':deflect.ion curve for the beam
indicates general yielding. Section .12 deals with the 'analysis at yield stage,
and an assumed distribution 'of . curvature which will take .into account this
difference between the theoretical and actual curvature distributions at yield
stage ,is discussed' .
. The difference between theoretical and measured curvatur'edistribu-
tionsat crushingstageisgreaterthan at either 'ofthepre'Vious two stages.
Figures 17 through distributions for three of the beams of
this .investigation.Atthisstage ,the theoretical curvature distribution
fails to explain the large deflections actually measured in tests of.reinforced
concrete beams. The difference between the yield moment and the erushingmoment
is small, and the area under ,the theoretical curvature distribution diagram of
'Fig .. 15c neglecting curvature within the stub yields considerably smaller
calculated deflect.ions at crushing Btage than those measured. The limiting
case wouldbe'the idealized moment-curvature -relationship shown in Fig. 20
where the yield and crushing moments are equal. Thismoment-curvature
relationship leads ,to ,the theoretical curvature distribution of Fig. 21, which
.yields the 'Same computed deflection'at crushing stage Hsthe computed yield
deflection. erushing 'curvatureis assumed to occur only at
-23-
one 'point following 'the :' idealize.d moment ,:"curvature relationship. This) of
course) does not agree with the. observedbeJ;lavior 'of a -reinforced concrete
beam where the crushing deflection is considerably greater t h a ~ the yield
deflection. The actual distribution of curvature is of the form shown in
Figs .. 17 through 19 ) the peak curvatures being distributed over some finite
length. Section 13 deals with analysis at crushing stageJand consideration
is given to. an equivalent distribution of curvature for 'useincomputing,the
crushing -stage deflection.
An examination of the theoretical analysis 'rev.ealsthe tacit
assumption that for ,the g:iven cross section there is a unique :relationship
between moment and curvature. Tb,isleads ,to the conclu.sionthat a prismatic
'reinforced concrete 'member 'havi;ng .constant reinforcementalong:its 'length
would have a uniform distribut,ion of curvature' in areg-ionof 'constant ,bending
moment) and, therefore constant s-trail1, at the top fiber . . Experimental ,measure-
mentson members with a ,constant moment regiort having a finite number of
cracks as shown, in Fig . 22a indicate maximum concrete strains at the top fiber
above the cTack and ':minimum values ,in between cracksasshown, in Fig . 22b.
The neutral'axis 'position also varies)beingpigherat the 'section of a crack
than between cracks. If curvature is defined as the :ratio of top f'iber con-
crete -strain to the depth 'ofthe 'neutral axis) the distribution of curvature
must vary as shown in ,Fig. 22c even withi;ntbe reg-ion of eonstantmoment.
This distribution.is quite different from the assumed uniform distribution of
curvaturewh'ich the, unique moment ... curvature -relationship would predict .
The validity of the unique moment-curvature relationship ,is dependent
on the assumption that strains'be distr-ibutedlinearlyover ,the section. This
assumption would be approached if there were an infinite number of 'flexural
cracks) but for actualmembeTs, having 'a fini tenumberof cracks the distribution
-,24-
of strains oyer thedepth -of.' thes'ect'ion ..is notlihear .. The error in analysis
of 'peams having -a constant moment regionis not very':large
since the 'assumption is satisfied on the average over the flexural span .
. However ) for beam -to -column connections). theconc.entrated angle . change occurring
at the joint produces a . large deflection at-crushing, ,making:this
error "in distribution ofstrains:rtluch more -serious .
. After yielding 'ofthetension steel) ,morec-racksoccur 'in the
vicinity.'of the joint and a .loss of 'bondhetweenthe concrete and the tension
steel results. If' the 'concrete-strain distribut.ionalong the top ,face were
plotted) it would . form of Fig. 23bat crushiri'g') the maximum value
being .,atthe . c-ri tical section . . The 'steel .strain distribution
the form of Fig. ,23c 'atthis 'same stage -of "loading, the Btrains
constant in the region where bond ,is lost toa large:extent. If the neutral
axis,.is:in the: position shown in Fig. 23aandstrainsare:linearly distributed
over the section,,' the .. curvature -maybe :computed from-the' concrete strain
cpO = E /kd'
c
or . from the steel strain
However) the actual distri'butionof curvature compatible with the
concrete ,is, that -shown ,by, :the dashed 'line in ,Fig, '$'3d
while the curvature distribution:fromsteel stra.ins is shown pythe solid
line of :the . same figure . Since the form of the ,two distributions is. diffe.rent)
..
it follows that they cannot be the same at one point;
,therefore) .the distribution .of strains is . not:linear 'overthe 'section in this
region of peak curvature as assumed. ,Th.e ,:sametotal angle change must be
satisf'iedby thesetwocurv,ature d:i,.stributions '. which that the 'area
underthe curves bethe same.
-25-
'rhe-computedcurvature'atthe critical section at crushing ,is
probably some intermediate value between the :real curvature from strain in the
concrete and that- from,-steel strain described Both of these distribu-
t ions show a _ II spread
ll
of some peak value in theimmedia te vicinity of the
critical section. The same angle change could be produced by a spread of the
computed curvature which'has the same area under the diagram. If the centroid
of this -area _-is approximately in the ,same position 'as :thatfor ,the real dis-
tribution) the :samedef'lection would be-calculated using this eCluivalent spread
of computed curvature rather ,thantheactu.;Ll distribu.tion _of curvatures from
concret'e or' steel strains. -The spread of this computed curvature as shown in
'Fig. 23d has been one of 'the 'objects of 'this investigation. If the length) x)
can 'be predicted by a consistent procedure and the curvatu+e computed in terms
of eross-se-ctional and-material properties of the-member) 'thenthiseCluivalent
distribut:ion maybe used to predict the angle change and deflection of the
member. ,This method. is 'comparable to us ingan eClui valen t stress b;Lockfor
compression ,in co;mputations instead _of the real
distribution which is actually 'rather 'uncertain. ,Section: 13 deals with
analysis-at crushing:stage.
Only the moment-curvature :relationship up to'thepointoffirst
crushing ,of the concre-qe has been considered, thus far) and most prior-work
has taken this point to represent the ultimate '$tage in the behavior 'of the
beam. -Actually) this is a grossly inaccurate concept -for 'memhers wi theven
moderate a;mountsof compression reinforcement and c-losed ties for stirrups )
whose deflections at ultimate are much greater -than the deflection sat.-'fir.st
crushing. Of course, the moment at .higher ,than the crushing
,moment . Estimating:this ,point ,in the behavior of a ;member -has been one of
the objects of this .study.
-26-
In order topredictthisstage 'in theload.-deflectionbehavior, we
must deal with a 'cross section'whichis different :from the original
member, and the real relationship between mQment and curvature at ultimate is
even more uncertain than at; crushing. It is Sin observed fact, however, that
almost all .the additional deflection beyondcrushing:isa result of curvature
which occurs in the immediate vicinity of ,stub. This leads one to
assume that the :real distribution might be approximated by an equivalent dis-
tributionconce.ntratedin this:region .for 'use inpred.,icting,theultimate
deflect ion. of .the :member. , The magnitude of the curvature in this ease must be
relat'ed to the cross se:ctio.n which remains after crushing has becomevery
severe. involves the 'properties of the 'confined conc.retewithinthe
closed stirrups, the outside the confinedcore being :severelycrushed
by this stage.
In this study) the ultimate curvature of the 'reducedsection
consisting of confined concrete and steel areas has been computed. An object
of the investigation was to arrive emp:irically at an ,eq,uivalent distribution
of this computed curvature which would predict the observed ultimate deflection
as olose1yaspossible. .The computed moment from assumed behavior 'of the
reduced section has also .bee.n:compared with the measured moment at ultimate.
11. Analysis at Cracking .stage
first s.ignificantbreak in the :load-def1ecti?n curve for a
reinforced .eoncrete flexural member 'occurswhen the concrete 'cracks .in the
tension zone . Prior to cracking, the beam has .the . stiffness of the uncracked
concrete 'S'ection, plus the transformed area of steel, and the behavior is
essentially 'elastic. when the load.isincreased until the computed extreme
fiber 'stress at the 'critical section .is . equal tot.he .modulus of'rupture of
the concrete, theconcre.te 'crac:ks, reduoing :the stiffness of the section.
-27-
It :is observed .int:eststhai;; ()ncethe crack occurs, it progresses well into
the member ) and. the change from theuncracked section stiffness to the lower
stiffness after 'cracking :is a Tatherabruptone. This change is observed
in the load.-deflection 'history of the 'member 'as a break in the curve 'from a
steep. slope initially to a flatter slope after ,the crack occurs. The pereentage
of steelin the member 'determines the 'relationship between the two slopes.
The continuous plott:errecords of'load versus deflectlontaken for the beams
tested .in this 'investigation, such as the one gtven in ;Fig. 6, show this change
in slope very clearly.
Theoretically, theipi tial stiffness .is a function of' the moment of
inertia . .of:the transformed sectionj but if only an estimate of the uncracked
stiffness is desired, the moment of'inertiaof the gross section gives a,. good
approxi7nation. ,This greatly' simplifies the computation sincei t does not
. involve the :shift 'in the neutral axis for unsyrnmetrieal arrangements of tension
and . compression stee,l. The curvature, cp, for the uncracked section ,is thus
'simply
where M = applied moment .
E ,= modulus ofelast ici ty of' concrete
c
I = moment of . inertia of the gross section.
Assuming:elasticbenavior up to this. 'point, the distribution of
curvature along :the span follows the moment diagram as shown .inFig. 24a. The
curvature within the stub may 'be neglected since the momentof'inertia of the
stub is much 'greater thanthat of the beam. . The deflection at midspan is
.then

A
'6 = (P/?EcI) .. \ /2)(2/3)
'. /:::. /
6E
c
I
where .is the distance from support to the face of the stub .
(1)
This expression' '6 and P prior' to cracking 'and g-i ves the
initial slope of the load-deflectioncurve. The load at 'which cracking occurs .
is taken as that for which the computedextreme fiber 'stress at the c-ritical
section :is :equal to the modulus of rupture of the concrete. Values of the
modulus 'ofruptureobtained from tests of control beams of eon crete used.in
the current program and from other tests 'reportedby Warnaruk (4) are pl.otted
versus the compressive.s.trength in Fig .. 25,' Theeraticnature of the valu'es'
from thistestisreadilyapparentfrdm the scatter, but the trend of the
data is well represented 'by 'Warwaruk I s expression
'. . . : 3000
f .-
.r 3 +12,000/f'
. '", .... c
(2)
where the values of 'both f -and f" 'are in :pounds 'per squareinch.
r c
Because of the wide Bcatterinindividual test values, the modulus
of rupture is assumed tobea .function of the cylinder 'strength,as g-ivenby
this 'expression, in the analysis for thecracking .load. Once the modulus of
rupture is known,theeracking :load can be computed as.follows:
.f Me
; r - 1
P
c.
r
3P .J.,
cr
bh
2
where h = overall depth of the' cross section.
In many cases, small shrinkage cracks 'existed at the corners of the
stub before any.'load was applied to the member ,and the erack at the critical
section tended to form ata slightlY'lowerload than predicted.
-29-
In summary J the cracking :load maybecorhputedfromEq .. 3, assuming
the modulus of rupture to be Telated to the cylinder 'strength according to the
expression ofEq. 2. The cracking 'deflection is given by. Eq. 1.
12. Analysis at Yield Stage
The secqnd significant break in the load-deflection curve for'a
reinforced concrete member occurs whep. the tension steel yields .At this
point the load becomes 'almost constant while deflection continues to increase.
The present stuCly .. is cOr).cernedlionly withunder""reinforced .beamsin which this
yield point is reached before the 'concrete strain is high enough to produce
crushing. The :yield point.is very importa,nt since i t 'repJ;"esents the boundary
betweenelc;1stic and. inelastic behav.ior ;for the load";de!'lection 'envelope of
thebeam.
The load and deformation atthe yield;point can be satisfactorily
'determined by use of the conventional "straight .line theoryll with slight
modifications. This. method re q.u, ires the following aSE?umptions:
'1) Lineardistribution of stra;i.nsover the depth. of the section.
2) Linear'stress-strain:relationship for 'concrete 'and known
modulus .ofelasticity.
3) Concrete carries no tension.
4) Known relationship and modulus of elasticity
for the steel.
The resulting:linear 'stress distribution for the Goncrete .is correct
only if' the maximum stress at the top fiber 'of'thebeamis less thana-bout
half thecompres si vestrength of the concrete. Theunder-reinforcedbeams
of this investigation satisfy this as su.mpt ion reasonably well since the steel
. percentages were well below the value which would be -requ;ired to produce a
balanced failure by simultaneously crushing :theconcrete andyielding:the
tensionsteel. Ifamo;reexactcalculationis req.uired tl1e real stress .... strain
curve for the . concrete .in thebeam must be known .
..:-30-
The s'traight :line theory:illay'be 'usedto ,locat:ethe position of the
neutral axis 'as follows:
I b I
s
d
'
E
C
--I!I_I--+------I---I--- "----r=-....4 --------- -A __ --'I!I_ T
Esy
StTain Stress
For ,the beam s.hown :W;'ith ,both tension and compressive reinforcement,
the depth to the neutral axis
J
.kd
J
according to the 'conventional straight
line theory. is ,found usingthe 'following ,eX!?ression for 'k:
k =V2[pn + p' (l-d-d
'
) (n-l)] + -[p' (n ... l) +pn]2 ... [pl (n-l) + 'pn]
where
.' p = "
p' = AI
$, '
n '= E IE
" .. 'BC
In these studies themodu;lusof,elastic.ity of concret-ehas heen assumed to
be a function of the compressive 'strength as 'follows:
30,000
6
where both ,E 'and f1 arein kB i .
iC C
(4)
The tens-ion's,teel st:rainat yield is .kn9wnfrom the stress-strain '
curve for thematerial. It ,is-necessary that this 'yi'eld ,strain be ,known as
accurately 'as possible,s-ince the yield curvature 'is very sensitive to errors
in its 'value . The 'curvature; cp J! is,simply
Y ,
-31-
.cp = E / Cd - .kd) ( 6)
y . sy
whereE . is the tension ste.e.l st:rain at yield. The expression given inEq,. 8
sy
may be used to estimate this 'v:alue for analysis.
The internal moment at the critical section at yield stage is easily
computed;once the axis. position is found and the magnitude and dis-
tribution -of strain have been established. If compression 'steel is present,
the strain in -it is
E' = cp ( kd -d I )
S Y
If this is lss'$thanthe yield strain, force in the steel, 0 I,
S
,is
e' = E'E A' .
s s s s
'If E 'is ,greater ,than the yield strain, then
s
Of :::: A'f"
s s y
Since the section must he 'ine<luilibrium, the total compressive
force musteClual the force in the 'tension :steel at -yield.
Thus
'and
or
T o
Af = G' + C
S Y s c
o =Af -0'
c q y s
The moment ,at the critical section at yield stage is then
M
y
(Af - A 'f" ) (d -.., ._kd) +A "f' (d-d' )
"S Y s s 3 - ,s s
For abeam without compression steel,the neutral axi$ 'may still.
he located by -Eq. 4, although the expression becomes 'Somewhat simpler -in this
case. The yield curvature ;may then be -found from 6. Equation 7 gives the
yield, moment, the :second term being -:zero When there is )10 compress ion steel.
-32-
.The . deflect.ion at yieldstage.may 'be 'computed .if we . know the
distribution of curvature along thespan in addition to the curvature at the
critical section given by Eq. 6. A typical measured curvature distribution
from concrete strain measurements 'is shown in Fig. 16. This distribution along
. thespanfollowsthetheoretical distribution at this stage as discussed in
SectionlO:except for thepeak curvature measured ,in the 2i ..... in. interval
I
adjacent to the stub ,face. . However) it is important to note that curvature
was also measured within tl:J.e colunm .stub) and this curvature makes a very
s.ignificantcontribution to the deflection. Assuming theyield curvature
to be distributed uniforrnlythroughout :the 12-in. width ,of the 'stub 'givesa
good approximation of the real dist,r-ibution; ; This distribution) as shown .in
Fig .. 24b ,was assumed by McCollister .:in his 'analysis for yield deflection and
has also been used in this study for beams having w = 12 in. . Figure 26 shows
.. - --'-- "--'-"- --'-
tliedistrI;'6ution assumed for, Ernst 'sbeamswhichhad different stub lengths.
Since the analysis of previous investigators (1 ,2: ) using 'conventional
straight line theoryhavecons.istently given .lower "predicted deflections at yield
than those measured in tests) it is not surprising that this was found to be the
case in analyzing the beams of thecurrentprogram. Since the measured dis-
trrbution of curvature seemed to ju.stiy" ,the ;e'luivalentform used in the
anaJ_ysisbyMcCollister and. in this a further -study was made of the
factors-wnich influence the :yield curvature computation.
The 'curvature at the critical section is given byEq. 6 as follows:
cp = E j(d - .kd)
y sy'
(6)
This expression is very . sensitive to the value of the strain) E ) .which is
.sy
assumed to correspond to yield. The position of the neutral axis is also
involved) and factors 'affectingthe computed value of'kd also should be
considered.
-33-
The yield strain commonly -assumed to hef /30,000 wheref is
y y
expressedinksi" and thestress--strain curve is idealized as follows :
f
f
Y
E 30)000 ksi
E = f /30 ,000 E
y y
Tensiont:estson :reinfore.ing :hars used ,in the curreht . program show two
significantd;ifferences from this :idealized curve .. .First, the :measuredmodulus
of elastiei ty'for the defonned bars based on nominal area .isconsistentlyless
than '30)000ksi) :moreoften b:eingabout 28)000 ksi. The continuous load-straih
curves taken !.in thes.e t-est.smay be 'used to 'estimate E", but a consideration of
s
leads to the cf'Gllowingrelationship:
:nominal E
s
actual E
s
actual A
nominalA
The'ratio of actual area .to 'nominal area:isprohably .,less than one for the
deformed bars) and is almost certainly'so where the 'defonnationsare ground
off at a seet.ion :to allow a-strain ,gage to be mounted on the bar. Thus the
nominal value of 'Efor thereinforcing: ba.r in would .he
s
expected to he less than actual value'for thematerial) and this 'was the
ease in thesetests.
The' s'econd . differenc'ebetween ,the :measured and idealized stress-'
-strain curves is 'in the :shapeoI' the curve near the. yieldpoint . The break
in the 'meas\Iredcurveat yield :is 'nGt always so sharp 'as that shown above 'forthe
idealizedcurve. .Most of the load-strain curves which were 'plotted continu-
ously.'in the tension tests of thecurrent'programhad the fonn shown ,below:
-34-
f
f-+----A---------------
y
E
sy
E
S
Errors resulting:from these differences' from the idealized curve
are not. particularly. large , but the trend of 'both .is to give larger strains at
yield load than 'predicted by the idealized stress-strain curve. Althoughthe
.yield strain makes li ttleor 'no difference in the computed momenta tyield)
the yield curvature isa direct function .of this strain) and .these differences
do have a significant .:Lnfluence on ,the computed yield deflections. . The yield
:point for thebeam was .determined from the curve as the point
where deflection contiuued to.increase with no increase in load (see Fig. 6).
In terrnsofthe real stress-strain curve for the tension "Steel) 'this occurs
when .the yieldstres sisreached and. the strain corresponding to this point
is E
sy
If the idealized 'curve is used) the yield strain). and thus the yield
deflection) .willbeunderestimatedbecause of the two factors mentioned above.
Strain -readings takenonthe . tension steel ata .:point :one' in .. but
from the colunm stub ,for beamsofthecurrenttest 'program always 'indicated
higher valuesthanf /30)000 Jatyield. In addition to the consideration of
y
the real stress;..strain :relationship ;for thetension steel, this might be
expected fromthe definition of
ll
yieldload" used here. The tension steel is
eJast.ic up to the point .ofimpendingyield,and a' somewhathigher-strain must
be present whengeneralyfelding .of thebeam is observed from the load
deflection curve.
For. 'members w.i th high steel. percentages) especiallywhere little or
no compression 'steel is'provided, the assumption of.a linear -stress .... strain
--35-
:relationship for the 'concrete with .the"" modulus of elasticity givenbyECJ.. 5
'may notb:e sat.isfied. The 'effect of m$.Y. be illustrated
strain and corresponding :stress distributions. shown below:
f
1_ c -,
__ 0
>kd /
.........
< jd
= A f
Esy tiS y
Strain $tress
:The trend of.this line theory is in the direction of
.curva;ture, Cj)y,abOve thevalue 'computed by $traight line
theory since the neutral axis down. .This error' is not likely to be
'significant if the )Jlaximum stress inthe.topfiberof the beam i$less than
about half the compressive strength of the concrete,and the modulus of
elastici tyJromECJ. .. 5is a reasonable representat ion of the slope of the stress-
-strain :relationshipforthe concrete, which ,is almost linear up to this. point.
Measured curvature distributions at yield in the current. program
consistently 'showed "Curvature values inthe'vicini:ty of the c-ritical
section than the computed.yield ourvatureusing'Conventional straight'line
theory ,to locate the neutral axis andassuming:theyield stra;Ln as f /30,000.
y
. Figure 16 :shows themeaqured curvature distribution beamq of
the current test program at yield. The factors discussed above which influence
the computed yield curvature 'all indicate that conventional straight
theory.'may g:ivevalueswhich are too low, and the measured curvature distribu-
tion shows this to be the case. -At the: same this . measured distribution
justif'ies the form of the eCJ.uivalent curvature distribution used by McCollister
'as an approximation to ,the distr-ibution,and this 'same form has been used
herein.
. A ,consistent . means .was sought. .to correct .systemat:ically the computed
yield curvature so that. it 'could "be used with.,theassUmed distribution .of
Fig. 24b to predict the yield'deflections .'9-ccurately. . The. following :expression
for determining the yield for use in Eq. 6 'Was found to give 'consistently
good re.sul ts
(8)
This '. expression .contains yield strainoommonly assumed .plus an additional
strain ,increment which ,represents a correction factor.Theyieldstrainfrom
Eel- 8is used in Eq. 6 to compute theJli-ield curvature) ahd the correction
thought of as correcting:for ,the differences between real and
idealized st,ress-strain.r.elat.ionships for ,the steel as diseussed above )as
well asf'orthee:rror'in computing ,theposition.of the neutral axis by con-
ventional straight line theory.
This expression1{ould not be appropria tefor 'beams wi thhigh q values,
'where the neutral axis 'positionmust he located as'suming -:thereal stress-strain
'curve for' the 'concrete .
. 'Ehe distribution of curvature -assu..med foranalys.is neglects the
,,'.'
"break in the' theoretical distribution 'atthe' point wheretheerac-kingmoment
exists in the beam. since the difference between these two distributions 'is
slight) as indicated in Figure 16. Usingtheassumed curvature distribution of
Fig. 24b) the .yield. cleflectionis
S: ,2 2
'24- (3w +12wi. + B. ,.)
wherecp .' isthe yield curvature from Eq,.-6 and y:(isth'e length of the column
y
stub:
. Tn':suri1:rilary-.,. 'the position of' the neutral axis at yield.maybe computed
. .
from cOnventIonal' 'straighi line the'orj be given by Eq, ,5. The
'0'
-37-
.yield curvature is computed from Eq. 6 where the tension steel yield strain
is found from E(].. 8. The moment at the critical section at yield 'is given
by Eq. 7 and the yield deflection byE(].. 9.
13. Analysis at Crushing Stage
The third significant stage of behavior 'for 'under-reinforced concrete
beams is the point at which the concrete crushes' . The term under-reinforced
is used to des cribernembers . in which the cracking and yield stages. prec'ede
erushing 'of the concrete. If enough tension st'eel is provided" . it is pass ible
to cause crushing 'ofthe 'concrete before the has yielded, result-
ing:in a very brittle type of failure . Such amemoer' is called over-reinforced.
This brittle type of failure is undesirable s-incethereis 'very'little -warning
when collapse is near , and also because very "1i ttle energy. is absorbed before
collapse. Most building 'codesdo not allow the use of over-reinforced members
or -re'luire a higher -faetor of safety if their use is allowed. The analysis of
crushing 'in this 'secti0n is concerned only wi thunder-reinforced members.
In order to analyze beams at this stage of 'behavior, some measurement
mustoe associat:ed wi thc'rushing 'of the concrete . Previous tests by numerous
investigators (4,5) have shown that alimiting 'strainis usually,. present at
crushing, regardless of the conc-retestrength. Test results (1,4,5) show that
this -.limi ting-strainis ,aboutO. 003 to 0'. 005 for 'concrete strengths below
, about 5000 psi. Since small changes have little influence on the computed
properties at this stage, ,thecrushing:strain has been assumed as a constant
value of 0.004 in this analysis. Measured concrete strains at this stage of
'behavior 'for 'beamstestedinthe current program support this 'assumption.
In addition to this assumption, the following must be known or
measured:
:1) Stress--strain :relationship for 'Steel
2) Properties of the concrete compressive stress block at this stage
Distribtuion of strains over the depth ofthe section
Concrete carries 'no tension
The stress-strain relationship for the steel is known from the tension
tests made on,pieces cut from the same.bars .. This stage follows yield in the
. tens ion steel and. it is possible that strain hardening ofthesteel may occur
'bei'orethe concrete crushes. Thus) complete stress-strain curves are .needed)
and the point at whichstrain hardening 'beginsis important in this analysis.
In the current test program" . complete stress-strain curves were
available for 'all of the . steel McCollister Teported stress-strain
curves usually to only slightly beyond. stra.in .hardening. .Previous invest iga-
tors (lJ6) .have used linear assumptions for the approximate shape of the stress-
straincurve beyondstrainhardeningJ. .the . range of accuracy is <:iuite
limited for the linear approximation since the real curve 'deviates from the
straight line .
. TheiNriter .hasdeveloped an e4pression which more closely
approxtmatestheshapeof the intermediate grade steel
bars for strains 'between strain hardening and ultimate. Since the 'Stress.-
strain curve has a characteristic: .. ' shape :in this region and the ratio of
ultimate -stress to yield stress varies only slightly for intermediate grade
steel" .a shape function 'Was arrived to apprqximate this part
of the curve .. Anaveragevalue :for the ratio of ultimate stress.to yield
stress for steel used.in the beams analyzed in this study was found to be 1.69"
the range being 1.52 to 1.86 The ultimate stress occurs at. a strain
of about 15 to 16 percent. A correction t:erm was added to the shape function
so that the expression would yield correct stress values both at the beginning
'of strain hardening :and at ultimate . The complete generalized stress-strain
curve for is shown in Fig .27. This curve was
-39-
incoTporatedlnto the computer program used in the analysis -at crushing-and
ultimate (Appendix B) .
The properties of the compressive stress block in the concrete at
crushing have been assumed as shown in the sketch below.
St-rain -
0.004
k d
c
I;)-L-c
Stress
f -Ok d
cu c
The total forc'e in the concrete -is expressed in terms of average stress at
crushing, f ) and acoef. fioient, k2' is used to its point of action.
cu -
Since the stress distribution is known.to have a slightly different
shape -for different concrete strengths, the average stress atcrushing
J
f J
cu
is expressed asa function of the compressive strength as measured by cylinder
-tests, f'. Numerous expreSSions but the following
c
:expression g-ivenby -Warwaruk (4) was used in the analysiS he;rein:
f
cu
where 1"1 and f are both in psi.
c cu
8 f
l x.10-4
0.' +
c
The point of action of the compressive force in the concrete is
assumed to be k2kcd from the top of the beam. There is probably a slight
variation of this constantk
2
with changing :shape of the stress block for
(10)
-different concrete strengths . The change is small) however, since the extreme
limits of this constant are 0.33 to 0.50 for a-triangle and a rectangle, re-
-spectively) as the -shape of the It is assumed in this analysis
that k2= 0.42 for -all values off J which is the same proc-edure followed by
-.40-
previous investigators (1,2 J 4). Any error in this constant would causeor;tly
a very small error'inthe moment arm at this stage, and
thus would have little effect on the computed moment.
The distribution of strains over the depth of the section is not
linear 'a t the c,rushing stage owing to the formation of inclined cracks in the
concrete. Strain measurements 'fJ;'om the current program indicate that the
strainsarelinear-ly distributed.in.the compression but the tension steel
strain ,is less than would be this distribution down
I . .
to the level of the tension:reinforcement.
Previous investigators have assumed the strain distribution to be'
linear 'over the depth of the section. If the tension steel is, in the yield
range at this stage , the' computed moment ';will", not be i;nfluenced' significantly
'by any error in the' strain assumed. However J if the steel is predicted to be
into the strain hardening -range when actually it is still within the yield
range, the computed moment' will be higherthan the measured moment. Linear
strain !distribution,over ,the:full depth.of' the section for the deeper beams of
, .
the current. program yielded computed moments at crush:i,ngwhlch were
5 to 10 'percent higherthan the 'measuredmoment at crushing. This errorwas
not as noticeable for the 'be:ams 'ha;v,ing d = 10 'in. since the tension ,steel was
usually. in the yield'rangeat this stage.
The distributi.9n of strain assumed for 'analysis 'and the corresponding
'stresses at erushipg'are qhoWn below.
, E = 0.004
.r4
c
.",,;
/
--Ir-. -k-
c
d------'---
0.42kd
c
f bk d
eu c
r::: : S s

T .= Af
. E:S . (:feu :from Eg. 10)
Strain Stress
-41-
"
: Strains are assumed to be linearly distributedin the compression zone, with
the limiting-strain at the top f'iber 'equal to 0.004 as discussed previously.
The strain at the level of the tension steel is assumed to be FE , where E is
s s
the steel strain .for alineardistribution of strains over the section. The
reduction factor, F, was taken as 0.85 f'or use in analysis after using other
values to cheek the.influenceot' this 'assumption on the computed moment.
Using this distribution of strains, .thepropertiesof steel from the
complete stress-.strain curve, and concrete properties intenns_:o:r.:;.effectivecon-
'cret:e strength, and k2' we can find the 'pos i tionof the neutral axis by assuming
trial values of kduntil the section .. is in equilibrium. The concrete is
c
assumed to carry no t:ension .. inthisanalysis. A program for the IBM 650
computer was written to perform .thiscomputation and to computeM, Cj), etc.,
at crushing. This'programis discussed,:in.Appendix B.
Having.located the.neutral axis by this 'procedure, the curvature
in the 'concreteat crushing :lS simply
E
. C
CPc ;= k d
c
0.004
k:d
c
(II)
The crushingmo;ment 1riay beeomputed directly once the position of
the neutral axis, kd,whichis consistentwiththe 'strain distribution and
c .
. material propertiesdescrihed above and resultsci.in eq:u,ilibrium of the cross
section, has beenlocated. The crlishingmomentis simply
M .= CA.f - AIf!) (d) (1 -.0.42k) + Arff (d-d')
cs s s s c s s .
(12)
The deflection at crushing involves the distribution of curvature
as'well as its .magni tude. Figures I 7 through 19 show': typical measured dis.-
tributions of curvature at c'rushingj the -relationship between the measured
distribution and the theoretical distribution was discussed in SectionlO.
Since this stage of 'behavior -is -not asprecisely defined as the two previous
-42-
stages, an equivalent distr:ibution, approximating the real distribution is
desirable for use in analysis. possible fo:rrf!.S of an equivalent ,dis-
tribution of curvature have been studied and are discussed below.
In the theoretical analysis discussed in Section 10, the spread
distance for peak values of curvature into the span at crushing is assluned
to be related to the moment as shown in Fig. At the point where the yield
moment is present, Section A, the curvature is theore.tically CPy and the dis-
tribution from this point out to the support, may be assumed to be linear. The
distance) Ct, from the face,of the colmnn stub out to this 'point is found from
geometry of the moment to be
M
ex = (1. - J...).
Me
Thus) the additional curvature that assumed yie,ld is spread over a
dist'ance from the center of the colmnn stub out to Section A, as shown in
Fig. 28. The curvature within this lel1gth above that at yield must account
for the additional deflection beyond yield) (6-6).
cy
The assumed curvature distribution shown in Fig. 24b was found to be
a good approximation for analysis of the deflection at yield, 6. At yield
y
the curvature at a distance ex out from the colu;mn stub slightly ,less than cp .
y
But the deflection produced adclitional between and
the support due to the increase in curvature Section A to CPy at, crushing 'is
negligible. The additional deflection beyond yield, (6 -6 ) must then be ,
c y
produced chiefly by the, curvature represented by the shaded area in Fig. 28
extending from the center of the stub out to Section A. If the shape of the
distribution within this region ,is reasonably constant at crushing stage) an
eCluivalent area can be found as a fraction of the area of the rectangle haviD:g
, '
a height of (cp' -cp ) and extent 8 Clual to' that of the shaded area. This equi va-
c y
lent area must account for the increase in deflection (6 -6 ). 'raking ,the
c y
-43-
centroid.of thisarea at the face of thecolunm 'stub 'should not involve any
significanterror . in view of the 'uncertainty of the measured deflection
i.tself ate-rushing. The increase in the crushing 'deflection beyond the yield
deflectionbythis method is then
W
6-8 = K . (cp . - -cp ) (- + ex)
c ycc y.2
where K is a shape fae-tor 'which detenuinesthe equivalent area.
c
Using measured deflec;tions" (.6.
c
-.6.
y
)' and computed values of (epc -cry)
and ex inEq, .. 13" the values for this shape factor, K , shown in Tables' 5 and 6
'c
werecomputedforthe beams of the current test program and McCollister's
beams. It is apparent that the shapefae-tor,K
c
' from .this anal;y-sis is 'not a
constant Hince thecomputed deflections range from 0.49
to 0.88 as shown: in these tables. A plot of K -versus a as.shown ,in Fig .. 29
c
shows 'a definite trend indicating that the' shape factor decreases with
. increasing a values , becoming :constant at about 0 . 50 for a = lOin. A linear
-approximation of this trend gives K
c
'asa function of ex as follows:
K= (0,9 - 0.04 a) 0.5
c
(14)
where a (1 - -M 1M ).
, .yc
The additional deflectionbeyond:yield, (.6.
c
'-8
y
)' was computed
from Eq,.13using :Kc 'as given by Eq. 14. This deflection was 'combined with
the computed yield deflection to obtain the crushing deflection. Chapter IV
contains a discussion of the results of this analysis for 'crushing 'deflect'ion
which has been designated as Method 1.
While this method has the advantage of relating thecomputed value
in a more or less rational manner to-the behavior 'at this 'Stage, it has the
disadvantage of requiring the computation of Cfl
y
' 8
y
' My' cpc,M
c
' a, and
(.6. '- -6.) in ordertofind b.. The possibility of erTorin the assumpti'ou
c y c
-44-
of the value of the shape factor, K , plus cUlllulative errors in the computed
. . .c .
quantities may offs'et the apparent advantage of this method to some extent.
A second approach was suggested by the observed of the measured
distributions of curvature in the current test program which showed that the
additional deflection at crushing beyond.yield is a result of peak curvatures
in the immediate vicinity of the faceof the column stub. The theoretical
value of the .yield and crushing' curvatures may "be computed from beam properties,
and the yield defiection may be predicted using the analysis of the previous
section. The additional deflection, 6. .-6. , may then be considered to result
c y
from 'an' equivalent rectangular curvature distribution having an ordinate of
(epe-CPy) and. spread avera length determined empirically from the measured
deflections. This leads to the 'equivalent distrtbution;of curvature at crush-
ing 'Shown as' the shaded area in Fig. 30a. Us ingthis d'istribution diagram, the
increase in deflection at crushing beyond the yield deflectiori is
6. - 6. = X I (cp ._ .. cp' ) ( )
c yc y
where x' is the spread length of this eCluivalent rectangular distribution.
Using measured (6.-6.), and computed values for (cp -cp )
c y c y
in Eq. 15, the computed spread length, x!, ranged from 5 .8t'o 10.0 in. for
these as 'shown in Tables 5 and 6.
In this case, the trend of the Xi values appeared to be related to
tension steel ratio, p.' The plot of x I versus p shown in Fig. ,31 indicates a
general trend for the value of x
f
to'decrease with increasing 'p,becoming
constant at 6.5 in. forp = 2%. A linear approximation of this trend gives Xi
as .afunction of p as follows:
x,
(10 l.75p) 6.5' in.
where p is the tension steel ratio in percent.
-.45-
The deflection beyond yield, (6-8 ), was computed from ECl. 15 using
c y
'x' as givenbyEq. 16, and this deflection was combined with the computed yield
deflection to obtain the crushing deflection. Chapter IV contains a discussion
of the results of this analysis which has been designated as Method 2.
It should be noted that Method 2 does not associate the spread length
1-7i th the moment . gradient as 1<TaS dane in Method 1. .If X' = K the t1'TO
c 2 I
methods are iclentical. Method 2 requires the compu-tation of cp , 6 , cp , and
. y yc
(6 _.6, ) in order to find 6. . The possibility of oversimplification in using
c y c
the distribution length, x', from Eq. 16 in the secondrnethod is something that
can be determined only as clata become available from additional tests.
The thirdmethocl ;i.nve;stigated utilized a simplification of Method 2
lhthe analysis of the data from the two test programs at the University of
Illinois. .AneCluivalentrectangular distr.ipution of the computed crushing
'curvature, cp , in thevicinitY'df the column stub accounts directly for the
c ,
measured crushing deflection ,in t.his 'rnet;hod, without i,nvolving :separate cal-cu-
lation of 6. Figure :30b'Showstheassumed distribution. In this case, the
y
spread length x' of Method.2.is increased to accountfor both the yield deflec-
tion and the ; increase in deflection beydndyield, (8
c
'-8
y
)' The additional
. ox" required to do this may 'be found by the procedure shown in Fig . 30.
Foy'the beams 'ofthe current prograrn'with w = 12 in. and = 66 in.
and for McCollister's beams with w 12 in. and 48 in.
8 I
X
The computed , values for the beams these two test programs are
x
given in Tables', 7 .and :08. These 8 r values are plotted versus tension steel
x
-46-
ratio)p, in Fig. 32. There is a definite trend for the additional length,o 1)
.. . x
to increase with increasingp up to about 2 percent, remainingessentially.
constant beyond this point. The linear approximation of this trend is as
follows:
When this additional length ofspreSid is combined with the x' values from
Eq. 16 .for Method 2) shown. graphically in Fig. 31, the total equivalent sprea.d
length x = x' :+- 5 f for.use in Method 3 is a constant as a result of the
x
. compensating trends of the two quantities. In this case thec'rushing deflec-
tion is solved directly from the equivalent curvature distribution of Fig. 3Gb
as
and since x Xl + '0
Xl
!:::. = xcp .
c c
10 in. for these beams, .
!:::. = 10rn
c "t'c
(17)
The computation of crushing deflection by Method 3 is certainly the
sirriplest possible approach. This method requires only cpand the assumed
c
distribution length, x, to c'ompute the crushing deflection directly. It should
be emphasized that this method is a simplification of Method 2 and might not be
as general in its application to beams with spans greatly different from those
tested in these programs. An analysis of Ernst's data indicates that it may
still be useful in approximating :thecTushing 'deflectionforvarying spans.
This is discussed in Chapter IV where the results of analysis for crushing
deflection by Method 3 are compared with measured values.
In summary, the analysis at crushing 'Stage incorporates the follo1;lTing
'assu.rnptions :
1) A limiting 'Strain of 0 . 004 at the top fiber is present at
cTushing'ofthe concrete.
4)
-47-
The :stress-strain curvefor'steel follows thecomp;Lete
generalized curve of Fig. 27.
The average concrete stress)f ) at crushing isa function of
the compressive 'strength as grV:en byEq. 10) and the coefficient
inlocating :the 'compressive force in the 'concrete
lso.42.
The strain in the tension steel is assumed to beF times the
'strainpredicted by extending -the .linear 'distribution of com-
pressive 'Strains -to this .level) where F is 0.85.
Tension carried by the concrete is neglected.
The posttionof the neutral axis is -found by taking trial values of
'k d until the forces acting ,onthes-ection eonsistent with the assumed strain
c
distribution and.material properties :result .inequilibrium of thecroB's section.
The crushing moment is givenbyEq. 12. The deflection at crushing
-is eomputedby one of the three following ::p:roeed,ures;
'-Methodl--The deflection in excess of the yield deflect.ion, j
c y
is g.fvenbyEq. 13. The distribution length, for 'peak curvatuJ;'eintothe span,
CX) is assumed to be-related t:o of, moment) and the shape
factor) K , is' assumed to be a -function of a as g-ivenby Eq. 14. ,The crushing
c .
defle ct ion is found by adding -this deflection) (8'..,,8 ), to the yield defTection
cy
,from Eq. 9.
Method 2--The deflection ,in excess of the yield deflection) (8 '--8 )"
c Y
is givenbyEq. 15 The spread length for the peak (cpc_cpy)' is
'assumed to be a function of the t:ensionst,eel ratio"p, as given by Eq .. 16.
The crushing . deflection is found by adding this 'deflection, (8. '--8. )) to the yield
c y
deflection :from Eq. 9.
Method 3--This 'methodconsidersthet'Otal crushing deflection to be
a result of angle change concentrated at the face of the stub, whichmaybe
estimated by a spread of thec,rushing curvature, 8. ,over -a length of '10 in:
c
The yield deflection ;in this case is apP:r:'oximated by increasing -the spread
-48-
length used in Method 2, and this method maybe thought o.f as a simplification
of that 'method. This method gives the crushing deflection directly from Eq. 17.
14.' Analysis at Ultimat,e Stage,
The fourth, stage of behavior fO'r an under,-reinforced concrete befu'11
is the point at which the 'maximum load is reached; in thisstuOy" this,point is
- designated as the ultimate stage. Grushing 'of the concrete is very extensive
by the time an, under-reinforced beam reaches its maximum 'capac i ty .
The:modes 'offailureatthis'point were discussed in Section 8, and the
analysis of this section has 'beenguided to a large extent by the observed
behavior 'at this stage.
Only 'beamshaving :closed stirrups, or ties,whichactas binders in
confining the 'concrete have considered ,in this analysis of ultimate
behavior. For' beamswi thout binders" the c.rushing 'stage discussed' in the
may! he considered effectively as stage.
The confined conc-retecore and compression steel when present act
t'Ogether 'in carrying ,the compressive force at ultimate. At this stage) the
concrete cover above the ties'and the compression steel is completely crushed.
This concrete may be -removed eaSily, leaving:thetop of the stirrups and the
compression steel exposed as shown in the 'photographs of'Fig.lO. It would
be purely f'icti tious to consider the original concrete cross section as
, '
effective in carrying :the ultimate moment in this case. computation of'the
force carried by the concrete, therefore, is based on the assUmed 'behavior of
the confined concrete core wi thin the closed stirrups or ties ," The confinement
influences both theeffecti ve -strength and" theul timate strain of the concrete .
Previous studies 'by Chan (7) on members failing by compression of
concrete having :differentamounts'of confinement resulted in expressions for
the ultimate properties of bound concrete. The ultima.te strain and the average
-49-
stress coefficient at ultimate were express.ed as functions of the ratio
between the volume of the binder and the volume of the bound concrete. This
concept has been extended by the writerto cover members having varying amounts
of longitudinal reinforcement, which was not a significant variable in Chanls
tests. A parameter, Pb' called the binding ratio, has been used in the expres-
sionsforthe ultimate properties of the bound concrete. It is defined as
Volume of stirrup + 0.1(D1/s)
Pb = volume of bound concrete
where D' diameter of compression steel
s spacing of closed rectangular stirrups.
Since confinement of the concrete is provided by both the stirrups
and the compreSSion steel, both of these must be included in the parameter
used in the expressions for the ultimate strain and the average stress at
ultimate for the bound concrete.
The following linear'expressions in terms of this binding ratio
were written on the basis of trends shown by Chan's (7) tests:
E ~ U = 0.011 + 0.20 Pb
where E is the ultimate strain in the concrete at the level of the
su
compreSSion steel, and
f ~ u = 0-. 0 +10 Pt ) f ~
where f' is the average stress at ultimate analogous to f at crushing.
cu cu
These expressions 'are probably oversimplifications of the 'real
(18)
relationshipsj however, they reflect the tTend in the values, and the results
of analYSis usingthem give consistent correlationswitn experimental results,
as discussed in Chapter :IV.
The stress-strain relationship for interrilediate grade steel was
discussed in Section 13, and the same complete generalized stress-strain curve
shown in Fig. 27 was used in the analysis at ultimate.
-50-
. The behaviarof the compression steel is an :Lnrportant factor here
since it is stressed into the yield range at ultimate for values ofp i Ip of
.1.0 or less. Some compression tests an reinforcing bars were made to justify
the use of the tension test to determine the stress-strain characteristics of
the compression steel. These tests are discussed in Appendix A. Tests on fixed
ended specimens with lengths ,:Qf 4to 8 in. between grips showed that the strain
was well into the yield range befare buckling occurred and that for the same
bar" strain hardening 'in campression initiates at slightly smaller strains
than in t:ens ian. In the analysis at ultimate stage" strain hardening of the
compression steel has been neglected even though it may have occurred in a
few cases. The yield stressframthe tension test 1IJas used as a measure of the
yield stress in campression" and the modulus af elasticity was taken as
30,,000 ksi in bath cases.
The distribution of strains in the compression zone was assumed to'
be linear as in previous stages. Because af the extensive crushingof the
concrete at this 'stage) it is impossible to measure the distribution afstrains
experimentally as was dane at earlier stages for some af the beams af the
current program. The strainin.the tension steel was assu.rned to be the value
found by projecting the linear distrIbution in thecampressian zone daiJlrn to'
the level of the tension steel. A :r.eduDt:ibn . factor ) F) would probably be
appropriate here as in the analysis at crushing. HO\,;ever), since the strain
at the level ofthe9ampression steel is much more variable at ultimate) this
'refinement daes not seem justified. Tension steel strains are veIl into the
strain hardening range at this stage and a reduction factor would have little
effect on the computed force in the tension steel. Due to the flat shape of
the stress..-strain curve as the strains 'approach ultL."TIate" a small change in
strainresul t:sin verY' Ii ttle change In .the computed stress.
-51-
In summary) the assumptions for 'analysis at ultimate are;
.1) Stress...,-strain .relationship for 'steel in tension follows the
generalized curve of Fig. 27.
2) Steel in compression has the same stress-strain relationship
as in tension, but strain hardening is neglected.
3) Strains are linearly distributed over the depth of the se.ction}
but the concrete cover over the' stirrups and comJ?ressionsteel
is neglected.
4):' The.' confined concrete core is assumed to behave in such a way
that its ultimate strain and effective strength may be
express'ed as functions of the binding :rat;io, Pb'
5) Concre.tecarries no tension.
Using the above assumptions, a suecessive trial approach was used
.to find the strain distribution which gave equilibrium of the cross section
similar to the procedure used atcrushing-stage. This analysis was made
using the IBM 650 computer yiththe program discussed in Appendix B. The
hinding ratio was computed from beam propertiesj and the ultimate strain at
the level of thecompressionstee.l, E ~ u ' and the -averag:econcrete stress for
the bound concrete at ultimate) f'! , were obtained from Eqs. 18 and 19,
cu
respect.i vely . The strain at the level of the compression 'steel determin,.es
the compressive force carried by the 'steel, strain hardening :being :neglected
in this analysis. Following the assumed linear 'Strain distribution, a given
position of the neutral axis corresponds to a particular 'strain distribution
from whiohtheresultingtensionandcompressionforcesmay be computed. The
solutionis found for the position of the neutral axis yielding tension and
compression :forces'whichare in equilibrium.
A
c
. '-52-.
C!=A
1
f'
S S Y
C =A f'
c ",'I.c Cli

! Es S s
Btrain Stress
In the analysis at this . stage only the confined concrete within the
stirrup ttfas cons'idered effective in carrying compression. The moment computa-
tion involves an additional assu.mp-tion'as to the point of action of the
resultant compressive :force 'il)c"the:concrete.The coefficientk
2
was assu.rned to
be 0 df..5 forthisanalysis.. Birtce<tb.eerror 'in this assumption will make only a
. .
slight difference ,in compu:ting :the moment) .this value
is . satisfactory .' 'The'forces are knoWn from the analysis described above) and
the ultimate moment is
nl
M = AIT! (d=d r) + (A f. - -A vfl ) (d-d i + -2" - 0. 45 a)
u s s . 'S s s s
(20)
The ultimate moment may also be 'computed from the empirical expression
used by McCollister (2)) but the approach deseribedaboveismore fundamental .
. It also has theadvantage'qfyieldingane-stimateofthe ultimate strain in the
steel and the ultima,te curvature for use,:-in computingtheultimate deflection .
. The analysis of ultimate deflection was 'approached in a 'manner
g:enerally analogous to that used in Method I for ,analysis, of crushing'".
deflections. The' spread of ;peak curvatures into the span is express'ed asa
function .ofthemoment-sas: .
where distance fromfaeeof '. stub to point in the where the
crushing moments,,: is present .
The peak curvature :La. spr.eadover -a distance extending :from'the
(21)
center 'of the'column'stub to the section where the crushing:momentis reached
-53-
f3.inches 'away from the stub face. Strain measurements taken after failure
from the punch marks on the tension steel for beams of the current test
program show that this distance gives a good estimate ofthelength'over which
the'peak steel strains are measured (Fig. 33). The curvature at ultimate) cpu)
is computed along with the ultimate moment on the basis of the reduced section
after . cover 'outside the confined core within the stirrups is lost. The ordinate
of computed.cpu and the spread length ( ~ +f3) enc-losean area which rationally
bounds the region of high curvature at ultimate; and a shape factor, K, .may
u
be determined from experimental data to estimate the equivalent area for 'use
in prediction of ultimate deflections.
The empir-:i.c'al'shape factor -K is -analogous toK , which was the shape
. u c
factor'used with Method 1 for crushing 'deflection analysis. However, the total
ultimatedef'lection .is computed directly as a function of the computed ultimate
curvature as was done in Method 3 forcomput.ing 'crushing deflections. The
ultimatedef'lection may then be expressed as follows:
(22)
'where: K
u
shape factor determined from test data given by Eq. 23.
cp
u
computed curvature of the reduced section at ultimate.
( ~ +f3)
spread length for . equivalent curvature distributiop.
Tables 9 and 10 givethe K v.alues computed fromEq. 22 using
u
::computed crushing and ultimate loads in determining the length f3 fromEq. 21,
and measured ultimate deflections. A plot of these K values versus p'/p is
u
. shown in Fig. 34. The points indicate a .slight trend for K to decrease with
u
increasing ;p'/p. There is considerable scatter as should be expected at this
stage of 'loading" especially for beams 'wi thout compression steel for which the
ultimate behavior is rather 'erratic. ,The foilowing expression is used in
analysis to approximate this trend:
K
u
, -54-
, pI ~
= (1.5 - 0.7 --, ) > 0.8
p ..
C.23\
. )
Almost all the points. 'lie' within a band + 20 percent from this line as shown
in Fig. 34. The paints shown as open circles represent beams having Hshear
ll
failures. They should be expected to fall below this band since their mode
of failure is different. These cases are discussed in Section 22 in connection
with the effect of transverse reinforcement.
IV. COMPARISON OF COMPUTED AND MEASURED RESULTS
15. General Remarks
The analyses at significant stages of behavior developed in the
previous chapter have been applied to the beams of the three test programs
described in Chapter 11. Since original test data were available only for the
beams tested in the current program and for those reported by McCollister, the
emphasis has been placed on these tests. Only the under-.reinforced beams from
Ernst's investigation have been analyzed.
Computed and measured values have been compared for the following
quantities: yield moment and deflection, crushing moment and deflection,
ultimate Irj.oment and deflection. The results are given in Tables 11 through 24
anq iQFigs. 35 th:rot!gh42. . In, the. tables ,the results are c o m p ~ r ~ d in te:rrns
of the ratio of measured to computed quantities, while thefigurespresent the
same data as plotsof measured versus computed quantities.
The effect of reversal of loading is discussed in Section 23 asi
separate consideration. Computed and measured results for members subjected
to this type of loading are compared in connection with that discussion rather
than in this chapter.
16. ' Tests at the University of Illinois
Yield Moments
The yield moments computed from Eq. 7 and the corresponding measured
yield moments forthe beams of the current test program and for McCollister's
beams are shown in Tables 11 and '12, respectively. The ratio of measured to
computed yield moment for 'beams of the current program had a mean of1.03 with
-55-
a range from 1. 00 to 1.08 as shown in Table 11. For the beams reported by
McCollister the mean was 1.03 with a range of 0.95 to 1.09 aE! in Ta-ble 12.
This good correlation measured and computed yield m.oments) a.s Sh01AIll by
the plot of'_ Fig. 35 for all beams) is a convincing demonstr9,tion of" the u.seful".
ness of the straight line theory in predicting this quantity' for under-
reinrorced beams. Considering the wide range of 'variables involved in these
programs) the sim.ilari ty between the mean va.lues and the ranges of values
for the ratio of measured to computed yield moment is striking. The dead load
moment1ATasincluded in the measured. values of yield mom.ent for the beams
of both programs.
Yield Deflections
The:continuous load=deflectioncurves ta.ken for the beams of the
current test program were used to determine the yield deflection of the -beam.
Load-deflectiou'curvesto an expanded scale for the early porti.on of the test
for all beams are given in Appendix C) and the cOITl-,Puted yield point is mark.ed
on these curves for comparison.
The yield deflections. -were computed from Eqo 9. Irhe mean of the
ratios or measured. to computed yield deflections 1-ras 1.02 for the beams of the
currentprogra;m) and. the range 'was O. 9
h
to 1.10 as Sh01-ill in Table 11. For the
hearns reported by McCollister) the correspond.ing mean "['v-as also 1. 02 with a
range or 0.93 to 1.13 as shovn in Table 1:2. The analys is f'or yield.
inyol yes the assumed distribution of curvature shm"rnin Fig. 24b and the
computed yieldcurya.ture usingslightly'rnodified st.raight line theory. The
commonly assumed strain in the tension steel was inc.reased a correction
factor>as discussed in Section 12. Consistently good resu1te. '\ilTere obtainEd
by'thiB- analysis for the beams of these t'lATO programs having a wiele range or
--57-
variables. Figure 36 shows 'aplot of measured versus computed yield
.deflections 'for 'all the beams.
Crushing 'Moments
Computed and measured cTushing moments for the beams of the current
program and those for "McCollister's beams are -given in Tables 13 and 14. The
crushing:momentswere computed from Eq,. 12. This analysis assumes the crushing
strain to be 0.004 for all members) regardless of concrete strength. The
average concrete stress) f ,is taken to be a function of the concrete
cu
strength as given by Eq,. 10) and a constant value ofk
2
= 0.42 is assumed. A
reduction factor) F = 0 .. 85, is applied to the tension steel strain obtained
fora linear -strain distribution in this analysis . The generalized stress-
strain curve of Fig . 27 is assumed to hold for all the steel reinforc'ement.
For the beams of the current progra:.m) the ratio of measured to computed
crushing moment as shown ,in Table 13 varied from 0.95 to 1.05, with a mean of
0.99. This ratio varied from 0.94 to 1.06 for McCollister's beams, with a
mean of '1.00 as shown ,in Table 14. The consistency of the 'resultsobtained
from this analysis is.obvious from the plot of measured versus computed crush-
ing :momentsin Fig. 37.
Crushing Deflections
Measured crushing deflections and crushing deflections computed by
Methods 1, 2, and 3 are given .in Tables 15 through 20 for the beams of the
current test program and for McCollister's beams.' This stage of behavior
'corresponds to the first evidence of crushing as measured by reversal of the
reading 'on'cone' of the strain ,gages mounted on the concrete nearest the critical
section or by visual observation. The measured results themselves at this
stage 'arenot as :preciseas at yield since crushing 'actually occurs rather
-gradually and no break in the load-def'lectiori .response of the member 'is noted
at this point. Readings are given to two decimal places in the tables only
because in most cases they correspond to load points in the individual tests
1-There the load and deflection were measured to this accuracy, and not because
the point is defined with this accuracy. For purposes of comparison, the
computed values from the three methods of analysis developed .in Section 13 and
the ratios of measured to computed values shown in Tables 15 through 20 have
been carried to two decimal places also.
The assumptions gJven above in connection with crushing moment apply
to the deflection computation also since the crushing curvature is computed
along with the crushing -moment. In addition, each of' the three methods
incorporates a different assumption for the :equivalent distributione,6f curva-
ture at crushing. These are summarized at the end of Section 13. The three
methods givewell distributed results which are within + 15 percent of the
measured values as shown in Tables 19 and 20. IJ:!he plots of measured versus
computed deflections are given separately for 'each of the three methods in
Figs. 38, 39, and 40. For the beams of these two testprogra.ms) it is obvious
that the simplified procedure of Method 3 gives almost as good an estimate of
the crushing deflection as the other more complex methods. However, the
third method is an approximation of Method 2
J
and the Xi. spread length for
(6 -6 ) and 6 I additional for 6 compensate to give the constant 10 in.
c y. x y
equivalent spread length for these members. For spans greatly different from
these) especially f'orverylong :spans) this might not be the case. Methodsl
and 2 are, perhaps) ,more general in their application although they are more
complex.
Consid.eringthe significanc-e of crushing as interpreted in this
study, the accuracy of thee-rushing deflection analysis (.:t: 15 percent) is
-59-
te satisfactory. The measured values may not be accurate to more than
0.1 in.) and first crushing as measured by reversal of a'strain gage
very minor damage to the member (Fig. 8). The test log and photographs
indicate only -localized surface spallingat the top surface adjacent to the
stub at this point. Since the first gage was approximately 1 in. from the
face of the stub) its reading -reversed as soon as the spallinghad spread to
this point) indicating crushing of the concrete beneath some part of the gagf=.
The pattern of crushing'whichdevelopsand the position of the first gage out
from the stub both influence the interpretation o:f crushing deflection.
The predicted and measured crushing loads and deflections are marked
on the measured load-deflection curves for individual members which are given
in Appendix C. These curves emphasize the large reserve beyond this point
before ultimate is reached for under-reinforced concrete members. It is also
obvious that this point is as far as the load-carrying
behavior 'is concerned.
Ultimate Moments
The ultimate moments computed from 20 and the corresponding
measured values are shown in'Table 21 for beamsof the current program and
for McCollister' sbeams. The ratios of measured to computed values for the
ultimate moment are also given) the mean being 1.00 with a range of '0.89 to
1.11. The analysis at this stage is based on the assumed behavior of the
beam after concrete in the compression zone outside the stirrups has crushed.
This involves an assumed strength of confined concrete within the core as well
as an assumed strain at crushing 'of this concrete and buckling of thecompres-
-sion steel. The steel is assumed to follow the generalized stress-strain
curve of Fig. 27j but strain-hardening of the compression steel is neglected.
-60-
For this 'stage of 'behavior) the results of the analysis are very good in view
,of the extent of c-rushing 'and the high strain levels in the tension steel.
Figure 41 gives a plot of measured versus computed ultimate moments for all
the "beams.
Ultimate Deflections
Computed and measured ultimate deflections are given in Ta'b1e 22 for
the beams of the current program and for McCollister's beams. The deflections
at this stage were computed from Eq. 22. Four of the beams of the current
program failed ,in a manner involving shear as noted in Table 22 (see photo-
graphs of Figs. 12 and 13). The ratios of measured to computed ultimate
deflections for these beams This is a result of failure in this
manner "before developing ,the ultimate flexural capacity with failure by crushing
of the confined concrete core and buckling of the compression steel as assumed
in the analysis. Figure 42 gives a plot of measured versus computed ultimate
deflections) the four points shown as open c,ircles representing ,the beams
failing 'in shear.
The ratio of measured to computed ultimate deflections varied from
O.74to 1.33 with a mean of 0,97 as shown in Table 22) excluding the four beams
noted as !I s hear failures." At this stage of behavior the ultimate deflection
for two supposedly identical "beams may differ considerably) and the analysis
cannot be expect'ed to be as 'accurate as at previous stages. In view' of this
and ,the extensive damage to the member 'at this stage) the correla-
tion "between measured and computed ultimate deflections is satisfa.ctory.
17. Ernst's Tests
The analysis of 'the previous chapter was developed primarily from
tests at the University of Illinois for which complete test data were available.
-.61-
When it.isapplied to the beams :reported by Ernst in which the stub length
wasamajor'variable) minor 'changesmust he made to take this additional
variable into account in the analysis. These changes are discussed below
. together with the results of the analysis. Only the under-reinforced beams
of Groups 2 and 3 of Ernst I s test program wereanalyz.ed. Group 2 beams
contained 3 percent tension steel and Group 3 contained '1 percent tension steel.
Five beams of each-group were tested at a slow rate and five at a fast rate.
Yield Moments
Computed and measured yield moments for Ernst's beams are compared
in Table 23. The measured yield moment-sare consistently slightly higher than
,the computed values . However J the maximum difference, is only 6perc-ent for
the slow tests and 9 percent for the fasttests. The mean: .ratio of measured
.to computed yield moment for the four -g-roups ranged from 1.03 to 1.05 with an
. overall mean ,.rat io of' 1. 04. Thes e rat ios are within the range of values from
tests 'at the University:of 'Illinois g-ivenin Tables 11 and 12.
No measured yield moments were reported by, Ernst for the fasttests
J
but the load-deflection curves forthese members were g'i ven . The"' yield
moments given, in Table 23 for these. members were ::computed using:loadswhich
were 'scaled from these 'curves) and the accuracy maybe less for these cases
than ,for the 'slowtests for which the yield moments were given.
Crushing Moments
Computed and measured crushing'moments for Ernst's beams are compared
in Table 23 . The 'ratios of measured to computed crushing :momentrange from
0.90 to 1.16 with an 'overall mean value of 1.06. This range of values is
slightly greater than the corresponding range for tests at the University of
Illinois given in Tables 13 and 14. The slow tests-of Group 3 beams had a
-62-
mean ratio ()f measured to computed crushing moment of 1.14) and the fast tests
of the same group had a mean ratio of 1.09. These mean values are somewhat
higher. than the mean values of-l.03 and 0.98 for the Group 2 slow and fast
tests) respectively.
The beams in Group 3 ) which had the highest ratio of measured to
computed crushing moment) have load-deflection curves which show a significant
increase in load w:i th increased deflection between yield and crushing (Fig. 43).
The measured moments given in Table 23 correspond to the load at Ernst's meas-
ured deflections) which are greater than the computed crushing 'deflections in
several cases for this group. Differences in the intierpretation of crushing
deflection will be discussed in more detail However) it should he
noted here that the lower computed moments combined with the somewhat smaller
computed d.eflections represent points on the measured curves) and possibly
indicate a difference in interpretation rather than a greater error'in computed
moment as compared to that found for the tests at the University of -Illinois.
Ultimate Moments
Computed and measured ultimate moments for Ernst's beams are shown
in Table 23. The ratios of measured to computed ultimate moment range from
0.95 to 1.12 with an overall mean value of 1.04. These ratios are within the
range of values from the tests at the University of Illinois.givenin Table 21.
The 'width of the confined core was assumed to be 4 in. for the
ultimate moment analysis of Ernst! s beams. The actual stirrup width was not
given) but the sketch of the beam cross section and photographs at ultimate
where the steel exposed indicate this to be a reasonable as sumption.
However) the results were not 'changed significantly when 4-1/2 in. was assumed
for the width of the core.
-63-
Yield Deflections
Measured and.computed deflections at yield for Brnst's beams are
given in Table 24. . Since the length of the stub was a major variable in these
tests, a s.light modification of the equivalent distribution of curvature at
yield from that given in Fig. 24b forthe Illinois beams havinga l2-in. long
stub seemed to be Teasonable . The yield curvature should not extend
all the way across the stub for the equivalent distribution in cases where the
stub length is morethanl2 in. The assumed distributions shown in Fig. 26
were used in the analysis of Ernst's beams to take this into account.
The measured yield deflections given in Table 24 are only approximate
since they were scaled from the load-deflection curves presented by Ern$t.
His tabulated yield deflections correspond, to yield at a .particular gage) and
in some cases it is obvious from the break in the load-deflection curve at a
smallerdeflectd.on that Hnother 'Section must have yielded previously. In this
study, .the yield point is defined in terms of the breaks in the load-deflection
curves shown in Fig. 43. Because of the uncertHinty of the measured value,
the comparison.ofmeasuredto computed yield deflection is made directly rather
thanin tenus of ratios as was done in the previous tables.
ForbeamsofGroup 2, the computed deflections of 0.34 tQ 0.38 in.
compare with measured values from the load-deflection curves of about 0.4 in.
in all cases. Forheams of Group 3, the analysis gives deflections of 0 . ~ 7
to 0.3.Q in. compared with measured values from the load-deflection curves of
about 0.3 in. in all cases. This agreement is quite satisfactory. Group: 2
beams contained 3 percent compression steel (3-#7 bars) with 0 . 037 percent
compression steel (2-#3 bars), thus the strain at the top fibeT in the concrete
at yield was higher thanforany of the beams of the current test program.
This high strain would indicate the strong possibility of non-linear stress
-.64-
distribution in the concrete) causing the straight line theory to under-
estimate the yield curvature as discussed in Section 12 and thus give lower
computed yield deflections than those measured. In spite of this error in
assumed behavior) the analysis gives a good estLmate of the yield deflection.
For the beams of Group 2" the error is in the direction which would be
expected from the discussion above. The beams of Group 3 contained only about
1 percent tension steel with the same compression steel as Group 2 beams. As
a result" the compressive strain in the top fiber is in the same order as for
the beams of the current program. For this group, the analysiS gives excellent
results since the straight line theory is more appropriate ..
Crushing Deflections
The measured crushing deflection is ~ u i t e sensitive to the empirical
criterion used to define this stage of behavior. The computed values compare
very favorably with measured values reported by Ernst as shown in Table 24,
with only a few exceptions. The trend of computed crushing deflection follows
. . '
the measured values conSistently, indicating that the approach is a d e ~ u a t e to
indlude the additional variable of stub length. For Method 1 analYSiS" the
w '
term ('2 + a) in Eq. 13 was modified to (4.8 + O.lw + a) . This express ion
accouritsfor the trend of effect of stub length variation" and yields (6 + a)
for the l2-in. stubs like those in the Illinois tests. The length is taken
to the face of the column stub and varies from 51 in. to 36 in. in ErnstFs
beams ~
Some of the differences between Ernst1s tests and those of the
current program which were used as a basis for developing the empirical
coefficients in the analysis are discussed belo1;lT.
Ernst used four different arrangements of concrete gages. For the
beams' of GrOups 2 and 3, gages were mounted on the vertical faces of the beam
-.65-
-at a point 3/4 in. down arid 3/4 in. out from the column face. The beams of
Group 3 had additional gages at the center of the top face of the beam and
3/4, in. out from the column face. Other arrangements were used for ,the fast
tests. Ernst indicates that gages on the top surface 3/4 in. and,2-3/4 in.
away showed no "consistentor 'Significant strain differences." However, there
may well he a difference in lIcrushing deflection" as measured by gages at
these different positions, although it may not be consistent since it depends
'entirely upon the crushing 'pattern which develops, and tb,e definition of
,T! crushing" as applied to this point in the load-deflection history of a member.
Whether this difference is significant is dependent upon the degree of
prec,ision desired in estimating this stage of behavior. ,The writer agrees,
in general, with Ernst that difference is not significant, especially con-
sideringthe large :reserve capacity 'for deformation of these members beyond
this point.
r;I:he analysis 'from study was developed using a of
crushing as the point at which a ga"ge located approximately one inch away from
the stub face on the top surface began to show Teversal (Fig. 7). This
corresponded to visually observed surface spalling ,in this region of the
member (Fig. 8), and was usually noted as suchan the test log. Differences
of more than .20percent from this value should not be unexpected for a
different definition which may involve more or less severe crushing.
Reversal of strain ,reading on a gage was used to define crushing
by both Ernst and the writer,but it is difficult to relate this to a
precise deflection. Usually crushing spreads gradually, affectipg only a
part of the gage at :f'irst, .thencausingthe strain to decrease when it has
spread sufficiently to affect more of the concrete to which the gage :;Ls
bonded. In this study, all cTUshingdeflections were determined from plot'S
-66-
of strain ,measured by,the gage, on the concrete 1 in. out from the stub versus
the midspan deflection. This plot is much more useful than the load versus
concrete str,ain curve since the load changes very Slo'\lTly and is not a very
sensitive measure of the beg-inning of crushing as defined here. These two
plots may be compared as shown belovo
L

Point A
E
,c
p
L
c
Deflection cor-
respond.ing to P
c
E
C
E
The plot of deflection-strain gives a lower "bound value 'if point A
in the sketch above is used as an indication of first crushing. The load-
strain curve will yie'ld a higher deflection before complete reversal is noted.
It should be note'd 'tha twhere there is a significant difference the
measured and-computed values for 'Ernst ITs beams in Table 24" the analysis gives
a conservativeesti.11late of the crushing deflection. This difference could
resu'ltfrom the use of the different plots given above in the analysis of test
data to determine thispointj that is" the University of Illinois tests rn..ay
have been interpreted more conservatively from the deflect'ion versus strain
plot than Ernst's tests from the load versus strain plots.
Ultimate Deflections
The measured and computed ultimate deflections for :Ernst!s beams
are shown in Table 24. The measured ulti.11late deflections are the valuesgiven
in Table 9 of Reference 3" except for those beaml? :for'N'hich the maximum load
-67-
-or a load only slightly' less than the maximum was sustained to a deflection
greater than the tabulated value before a sudden drop in load occurred. For
these cases) the deflection given in Table 24 corresponds to the maximum
deflection to which this peak load was carried as determined from the curves
in Fig. 43. Considering the uncertainties in the analysis at this stage) the
agreement between measured and computed ultimate deflections is satisfactory.
The same correction for length of stub other than 12 in. was used
here as at crushing. The term. was modified to become (4.8 +O.lw + 13)
inEq,. 22.
Rate.ofLoading
Although the rate of loading was a variable in Ernst's tests} it had
little influence on the obs-ervedbehaviorof the beams exeept for
deflections reported for some of the beams in Group 3. The analysis gives a
generally good estimate of the'reported crushing deflection for most of these
beams) but the differences were somewhat greater for some of the slow tests.
However) even these tests were completed in a shorter'period than the tests at
the University of Illinois. Thus, the crushing deflection predicted by the
analysis herein is a 'lower bound as discus$ed previously, even for tests at
considerably faster -rates than the tests from which it was developed.
18. Summary
In sunrrnary) the comparisons of measured and computed results p;r-esented
in this chapter indicate that the behavior of under--reinforced concrete members
may be predicted means of the analysis which was developed in
Chapter III. The three test. programs included in the study covered a wide
range of variables,and enough members were included to give some indication
of the scatter 'in the correlation between measured and computed values of moment
-68-
-and deflection w'hich might be expected at the v8..riou8 stages. While the most
uncertain predicted quantity is the ultimate deflection.for a member) the
analysis does give a goodestim3.teof the-reserve capacity beyond first
crushing. Considering-the use which is made of this. value the accuracy is
_ sa-tisfactory for predicting -this stage of behavior.
V. EFFECT OF VARIABLES
19. General Remarks
Beams from three different test programs have been included in this
study to give as wide a range of variables as possible. These test programs
were described in Section 5. McCollister and Ernst have discussed the effect
of the major variables included in their investigations, and their discussions
will not be repeated herein except as they pertain to variables studied in
the current test program.
The concrete strength was essentially constant for the current test
program, but in McCollister's tests it was a major variable, and the reader is
referred to his discussion of its influence on behavior. The area of tension
steel was constant for beams of the current test program and the tension steel
ratio,p, varied only as a secondary effect with depth. In McCollister's tests,
an extremely wide range of variation in the tension steel ratio with essentially
constant depth was studied. The effect of depth is discussed in Section 20
below, but the reader is Teferred to McCollister's discussion of the effect of
large variations in the tension steel ratio. The effect of compression steel
is discussed in Section 21 below. This was also a major variable and was dis-
cussed in McCollister! s study. In the current test program, the amount of
compression steel was varied for beams having each of the three depths studiedj
thus the . interrelation of the two variables is included in the discussion
herein, of the same observations as given by McCollister for
the single depth in his test program.
Several -of McCollister I s beams were subjected to reversal of loading,
but none were carried through several cycles of repeated and reversed loadirig
-69-
-70-
as -was done with three of the beams of' the current test program. Section 23
belmv contains a discussion of the effect of repeated and reversed loading on the
behavior of these beams. Their behavior is compared with that of companion
beams which were subjected only to repeated loading to failure in one direction)
and measured values of moment and deflection are compared with computed values
at the significant stages of behavior for these members.
Ernst's tests involved length of column stub and ra.te of loading -as
major variables. The reader is referred to Reference (3) fora discussion of
the eff'ect of these variables. In Section 22 below the inf1.uence of' length of
stub is d-iscussed in connection with the effect of variation in transverse
reinf'orcement in Ernst's tests. This section also contains a discussion of
the effect of variation in transverse reinforcement for beams of _the current
program.
20. Ef'fect of Depth
Depth was made a major variable in the current test program) at
least partly) because it had not been a variable in the previous two studies.
In spite of this" the spread of plastic hinging at crushing was associated
with depth by McCollister because it happened to give reasonable results for
the depth which wasinvestiga.ted" nominally 10 in. Ernst made no attempt to
predict the capacity for plastic hinging in his study ; he s-imply reported what
was measured.
The beams of the -current test.program had three different depths:
10) 14" and 18 in. Within each depth" three different compression steel areas
were used while the tension steel area was kept constant for all beams. From
this test program) the effect of' depth may be studied for beams wi th each of
the three :ratios of compression to tension steel areas; 0" 0.56) and 1.0.
-71-
Figure 44 shows the load deflection curves for -beams having no
eompression steel (p 1/ p = 0) but with depths of -10) 14) and 18 in. All the
beams were provided with No.3 closed stirrups at 6-in. centers. Concrete
-strength, f"f J and yield point of tension steel, f , varied slightly as shown.
- c y
The increase in yield load is a direct result of the increase in depth since
the area of tension steel was constant (2-#8 bars). Stiffness also increases
with depth -resulting in smaller yield deflections far the deeper members. The
increase in load after yield is primarily a -' resul tof strain hardening of the
tension steel. Between yield and c-rushing, the internal resisting moment arm
increases slightly, producing a slight increase in moment. Beyond crushing)
however J the moment arm dec-reases gradually, butthestraiil hardening of the
t-ensionsteel more than offsets thiseffeet. Thus the moment continues to
increase untilerushing -is severe produce failure of the member.
The effect of depth on ultimate deflection for these bearnswithout
compression steel is seen -clearly in Fig. 44, the ultimate deflection increasing
with depth. The c-rushing deflec-tion is li ttleaffected by depth, as indicated
on the s_arne figure. Thus ,the ratio of measured ultimate to erushingdeflec-
t ions (.6 /.6) varies from about 2 to 5 for the 10-, beams: respec,ti vely in
u c
this series. There is some indication that the ultimate deflection for the
deepest -bearnsmay be rather -erratic. Considerably- different ultimate deflec-
tions were measured for Beams J-4 and J-9 withd 18 in., even though the
beams were almost identical. However, Beams J-l and J-ll with d = 10 in.
showed very similar behavior. The higher -load a tul tiluatefor J-l resulted
from higher -fland f as well as initiation of strain hardening in the tension
c_ y
steel at a much-lower strain,
The effect of depth forbearns having two No. 6 bars as compression
steel and two No. 8 bars as tension-steel (pl/p = 0.56) is shown by the
-72-
load-deflection curves in Fig. 45. In this case, only one beam was tested
with -each of t..he three depths. The properties of these beams are on the
figure. The trends 'reflectedby these load-deflection curves are very similar
to those of Fig. 44 for beams vithout compression steel. Yield load increases
with depth and yield deflection decreases, the stiffness being greater for the
deeper beams. The ultimate deflection increases with depth but the crushing
deflection is al.most exactly the same for all three beams. It is obvious that
theul timate is considerably by the addition of compres-
sion steel" and the following section deals with the effect of compression steel
a.s a separate consideration. In this case, the ratio of measured ultimat-e to
crushing 'deflection (.6. /.6. ) varies from about 8 to 12 as the depth increases
u c
from 10 in. to 18 in. This represents a ratio forthelB-in. beams which is
1-1/2t1mes that for the la-in. beams. For the beams 'Without compression
steel, the ratio of ultimate to crushing -deflection for the 18-in. beams was
about 2-1/2 the same ratio for the lO-in. beams. This indicates that
the influence of depth -on ultimate deflection becomes less as the p't/p ratio
increases from a to 0.56. This effect. is discussed in the following :section
in more detail.
The ultimate load for the beams shown in Fig. 45 was considerably
higher than the yield 'load in each case, as a result of strain hardening of
the tension steel. Beams J -2 and J -14 failed by hy_ckling 'of the. compression
steel and crushing 'of the confIned concrete core (Fig. 11), but Beam J'-5
failed by fracture of the tension st.eel as indicated on the figure. This is
a rather 'dramatic illustration of the fact that there isa.limit to the
increased ultimate deflection which can be achieved by increasing the depth.
further 'increase in depth" the steel vouldundoubtedlyfracture with
perhaps an even smaller ultimate deflection than that forJ-5 with d = 18 in.
-73-
Beam J-5 also illustrates the influence on the mode of failure of the crushing
pattern which develops in the concrete. The cover above the compression steel
was completely crushed and had been removed before ultimate forJ-5, but the
extent of crushing 'in the compression zone generally was not as severe as with
J -2 and J -14 in which the compression. steel buckled at failure. Since the
compression steel was well into the yield range at ultimate for these beams,
the extent of crushing which developed influenced the support offered by the
concrete to prevent plastic buckling 'of the compre$sion steel. Thus, it had
an influence on the ultimate deflection and, for a case as' nearly "balanced
u
as Beam J-5, the mode of failure.
Load-deflection curves are shown in Fig. 46 for beams witheq,ual
areas of tension and compression steel (p'/p = 1.0) and depths of la, 14, and
18 in. The cross-sectional and material properties for these bea.msare given
on the figure .. Yield load and yield deflection. again follow the same trend
as indicated in Figs. 44 and 45 for beams with p'/p == a andp'/p 0.56,
respectively; yield load increases with depth and yield deflection decreases
due to the greater stiffness of the deeper beams. The load increases beyond
yield, primarily because of strain hardening of the tension steel. The
crushingdeflections ;are almost identical for the beams having the three dif-
ferent depths. . However, . in this case , the ultimate deflection decreases as
the depth is increased which is exactly opposite from the trend'ofthe
previous two figures. Since other factors are essentially constant, this must
be related to the effect of increasingp'/p to 1.0. This is discussed in the
following -section dealing with the effect of compression steel.
The modes of failure for thelO-in. deep beams indicated that the
compression steel buckled and the confined concrete core crushed at the same
time to produce failure O.r the member . Beams J -6 and J -13 , however, failed as
-74-
a combined buckling 'of the compression steel and shearing movement along one
of the cracks 'at the critical section. This shear displacement is clearly
indicated by the shift of the original centerline scribed on the beam at the
beginning of the test in the photographs of Figs. 12a and 13a. The ultimate
deflection for these beams was und,oubtedly less for this type of failure than
it might otherwise have been. for:.a .. '. flexural failure (see discussion in
Section 22. Strains'measured in the tension steel were very close to ultimate
at the time of failure J and Hearn' would have':' failed by fracture of the
tension steel at only a slightly higher deflection if the shear failure had not
occurred. Beam J-5 with p'/p = 0.56 and d = lS in. did fail by fracture of
the tension steel.
The ratio of ultimate to crushing deflection for the beams shown
in Fig. 4.9 is about 14 for. d = 10 in.) but ford = 14 in. the ratio is only
about 11 and ford = lS in. only S. This is still a large reserve of ductility
beyond crushing. even for thedeepest beam (J-6), but it is importan-:t to note
that the 'reserve becomes less with increased depth where pl/p =1.0.
Figure 47 shows load-deflection curves for beams having lighter
stirrups than those for the beams includ.ed in the previous three figures.
BeamsJ-19 and J-21 had No.2 U .... stirrups at 6-in. centers and no' compression
steel. The depths were 14 in. and.1S in. ) respectively. The trend of these
t"VITO tests is the same as that indicated. in Fig. 44 for beamswith #3 closed
'stirrups at 6-in. centers, the ultimate deflection increasing with depth.
Beams J-20and J-22 had No.2 closed .stirrups at 6-in. centers and two No. S
bars for 'both tension and compression steel. The trend of these curves is
exactly the same as that of Fig. 46 where pi/p = 1.0 and No. 3 closed stirrups
were provided at 6-in. centers. Beams and J-22 were both "shear failures!!
in a sense) but the ductility was less for the deeper beam. The behavior of
-75-
these beams is discussed in Section 22 in connection with the effect of
variation in transverse -reinforeement.
21. Effect of Compression St'eel
The previous section concerned the effect of depth, but the effect
of compression steel was found to be interrelated. Forp'/p ratios. of 6 and
0.56, the ultimate deflection increased with depth. For 'p'/p = 1.0, however,
the ultimate deflection was found to deerease with depth. This indicates that
both d and.p'/p are parameters which are associated with the ultimate deflection.
The interrelation between these two parameters, dandp'/p, ;may be
tllustrated in the following way.
When (p'/p) = 0) we see from Fig. 44 that the -ratios of ultimate
to crushing defleetion arerelat'edas follows:
8
8 u
~ 5 ford 18 in.
8
}
(8
U
)d
c 18
thus
c
;: 25
8 b.
u
2 for d 10
'"
in.
(b.U)d
8
-
10
c c
When (p'/p) = 0.56, we see from Fig. 45 that the ratios of ultimate to erushing
'deflection are related as follows:
8
b. u
3' 12 ford 18 in.
b.
(8
U
)d'= 18 c
>
thus
c
~ 15
b.
b.
(b.U)d
.U
8 ford 10 in. 10
&
-
'c
c
When (p'/p) = 1.0, we see from Fig. 46 that the ratios of ultimate to crushing
deflect.ion a;re related as 'follows :
6
6 '"Llj
"-
a ford lS in.
-z;:-
/ u)
c
} thus
'A'd 18
c
0.6
6- I:::.
-
'u
= 13
ford 10 in.
I u)
6 'Z=- d 10
c c
Thus) the interrelationship between effect of depth and (p'/p) may
.'be illustrated by plotting these ratios in the
25
2.0
6
(6
U
)d
lS
15
.0/
C
1, \ u.
6
1.0
(6 U)'d
10
05 c
0
0 0.2 0.4 06 O.S 1.0
pf/p
Sincethecrushin,es deflections are almost the same for 'beams having
'each of the threep' /p ratios ).practicallythe same . values are obtained in
each case from ratios of ultimate deflections as were found using the ratios
of 6 /1:::. 'as shown above " ; the :ratio I:::. /8 is an indication of the
u c u c
'res-erve deformation capacity beyond crushing" and the beneficial effect of
compression steel is -reflected by the significant increase in the .magni tude
of this 'ratio :(orthe beams having p'/p = 0.56 or 1.0 compared with that for
'beams having pi /p = O.
The corresponding to t4oseshO'WIl above may also be computed
using the predicted deflectionsf1rom the anq.lysis -rather than ,those measured.
This was done for the beams 'of .the current program" and the points are shown
on the'plot above as open circles. The trend of these points generally follows
that of the points computed using measured deflections. At p 1/p = 1.0 the
-77-
theory predicts a slightly higher ratio than the measured value; but the
ultimate deflection for the beam with d = 18 in. (J-6) used for 'computing the
measured point may be low as a result of themode of failure involving -shear
for ,this beam) in contrast with the flexural mode of failure observed for the
other beams and assumed in the analysis.
The interrelationship of depth and p'/pindicated above is probably
related to the fact that both of these beam properties influence the moment
gradient at ultimate and the ratio Me/Muj andsince the spread of peak
curvature into the span is related to the moment gradient with {3 = (1 - M /M )
c u
in the analysis) both d and pi/p influence the ultimate deflection. The ratios
of deflections shown in the plot above are functions of depth) and both the
observed behavior and the .analysis indicate that the influence of depth on
ultimate deflection tends to decrease as.p '/pincreases. As pointed out above)
the ultimate deformation capacity is increased as a.resultof the addition of
compression steel regardless of depth.
The effect of compression steel maybe examined by comparing directly
the load-deflection curves for beams of the same depth as the ratio pi/piS
increased. For thecurrent t'estprogram) the t'ension steel area remained
constant while the amount 0f compression steel varied from p'/p =Oto 1.0.
Load .... deflection :'curves are compared in Figs . 48) 49) and 50 .forbeamshaving
depths of '10) .14) and 18 in. J respectively) and varying amounts of compression
steel such that the pI /p .ratios are 0) 0.56) and 1.0 .
Figure 48 shows the effect of compression steel on the measured
response for beams with d = 10 in. The most obvious trend is
the marked increase in ductility which results 'from the addition of compression
steel. The Yleldpoint load is influenced primarily by the yield point 0fthe
tension steel; but the concrete strength as well as the amount and properties
, -78-
of compression steel have some influence on the internal resisting moment ann,
and thus the yield point load. Beams J-l and J=ll with no compression steel
show slightly different yield loads. However) both the yield point of the
tension steel and the concrete str.ength are higher for J-l than J-Tl, and the
yield load for 'J-l is correspondingly higher. For ,Beams J-8 and J-17, with
equal amounts of tension and compression steel, the yield loads are practically
identical, the concrete strength being higher 'forJ .... 8 and the yield point of
the tension steel higher 'for 'J-17. The highest yield stress ,for tension steel
in this group was for Beam J -2 which had a pi /p ratio of 0 .56. The yield load
was correspondingly even though the 'concrete strength was
The erushing 'deflectionsvary only slightly for the beams with
d = 10 in. shown in Fig. 48. There is an increase in 6. with additional
c
compression steel due to the higher 'crushing 'curvature, but the change is very
small.
Beams J -1 and J -,11 with no compression steel showed considerable
ductility since the tension steel ratio was only about 2 percent and No. 3
closed stirrups at 6-in. centers were provided. The comparative shapes of
these two curves 'show the influence of strain hardening 'of the tension steel
rather clearly. For Beam J-ll, strain hardening initiated at a strain of
0.0157 and the increase in ultimate load above yield load for this beam was
slight. For Beam J-l, strain hardening at a strain of only 0.0077 and a
larger ,increase in the ultimate load above yield load was measured for this
beam than for 'J-ll. The ultimate load was carried at deflections of '2.5 and
,2.0 in. forJ-l andJ-ll, respectively, but each beam carried a load about
equal to the yield load at a deflection of ,4 in. Crushing 'of 'the concrete
reduced the internal resisting moment arm while strain hardening of the steel
-79-
increased the tension force, until crushing was severe enough to produce
failure with a rapid decrease inload.
The addition of compression steel had avery oeneficialinfluence on
the ductility of the oeams withd = 10 in. as shown oy Fig. 48. Addition of
compression steel while maintaining the other oeam properties the same as for
J-l and J-ll produced much larger ultimate deflections. The failure was a
result of comoined oucklingofthe compression steel and crushing of the con-
crete followed oy a sudden decrease in loadwith increasing deflection. When
the o earns were unloaded before this ultimate had oeen reached, the same load
was carried upon :reloading and the load-deflection curve continued just as though
the beam had oeenloaded Figure 6 is a typical plotter record of
load versus deflection showing this repeated loading. Section 23 is devoted to
a more complete discussion of the effect of repeated and reversed loading.
At ultimate 'load, the tension steel strain was close to ultimate for
the oeams withp I/p 1.0. Any further' increase in pI /p aoove this value
would have resulted in fracture of the tension steel without any significant
increase in ductility. Equal amounts of tension and compression steel seem;'
. to be a practical limit so far 'asaddi tion of compress ion steel to increas'e
ductilityis concerned. For'memoers suojected to reversal of 'load this is
ooviously the case.
Load-deflection curves are in Fig. 49 for 'oeams having d = 14
in. with varying amounts of compression steel. The comoined effects of lower
yield point of the tension steel and lower 'concrete strength account for the
lower yield load for Beam J -10. These effects compensate for Bleams J -13 and
J-14, the yieldloadoeingalmost identical for the two memoers .
. Only one oeam of each cross section was tested in this series with
d 14 in., out the same trends were ooserved as for d = 10 in. The addition
-80-
of compression steel increased both the ultimate load and. the ultimate
deflection asp'/pincreasedfrom 0 to 1.0. Failure of Beam J-IO occurred
when the confined concrete core within the closed stirrups crushed (Fig. 9).
Combined buckling of the compression steel and crushing of the concrete were
the causes of failure for Beams J-13 and J-14. The bars buckled outward
over a 6-in. length between stirrups for Beam J-14as 'shown in the photograph
of Fig. lIb. In the case of Beam J-13" the bars buckled closer to the stub
with some indication of shear involved in the failure as shown in the photo-
'graph of Fig. 13a.
Load-'deflection curves are given in Fig. 50 for beams having varying
amounts of compression steel and d = 18 in. The yield loads for the different
beams of this series were almost the same. Slight 'variations in yield points
of tension steel and concrete strengths account for the differences.
Crushing deflectionsvaried.only slightly just as for the other groups. of
'beamswithd = 10 and 14 in. However" the over-all pattern of this group of
curves does not appear to be as consistent as those shown in Figs. 48 and 49.
The different modes of failure account .forthis to a large extent. The
behavior of Beam J -4 was more symmetrical than that for Beam J ... 9" accounting
'forsorne of the large difference in ultimate deflection for .these two beams
without compression steel. Beams J-5 and J-6 present an apparent. paradox in
their observed ultimate deflections as compared with beams of'the series with
d = 10 'and:14 in. However" Beam J-6 may have failed slightly prematurely when
compared with other flexural failures" although its flexural capacity probably
was almost reached since the tension steel strains were close to ultimate when
the II shear failurell,occurred as mentioned previously. The ductility ofJ-5
may not be completely typical since the concrete :crushing 'was never as severe
as with most of the other beams at ultimate. This was the only beam in the
-81-
whole group which actually failed by fracture-of the tension steel although
others were very close to failing in this manner. Considering these factors,
the curves of Fig. 50 seem to indicate that the beams with d = 18 in. are not as
much benefitted by the addition of compression steel as are beams with
d = lO andl4 in.
The beneficial effect of compression steel on ductility is observed
consistently in the comparison of load-deflection curves for all three depths
tested. Probably the surest way to add increased ductility tq reinforced
concrete members is to add compression reinforcement combined with closed
stirrups which serve as binders for the confined concrete core and restrain
the compression steel against buckling-in addition to carrying shear. The
analysis consistently shows the increased deformation capacity which results
'from these factors.
22. Effect of Transverse Reinforeement
The three test programs utilized in this study give onlY'limited
experimental data which may be used to study the effect of transverse
reinforeement. ,Most of the beams tested at the University of Illinois had
either #3 stirrups at 6-in. -.centers or a very close e<1uivalent, as shown in
Tables land 2. Ernst's Group 2 beams had two #3 stirrups at 4 to 6-in. centers,
and his Group 3 beams had one #3 stirrup at 4-1/2 to 5 : - i n ~ centers as shown in
Table 3. In the tests by McCollister and by Ernst, the amount of transverse
reinforcement _pro'vided was intended to prevent any possibility of shear failures
and was not one of the primary variables studied.-
In the current program, the stirrup size and spacing 'was constant
in all but four of the beams tested. The four beams in this series (J-19
through J-22) contained #2 stirrups at 6-in. centers in order that their behavior
might be compared with companion beams which had #3 stirrups at the same spacing.
-82-
The influence of the variation in. transverse reinforcement on the
behavior as observed from tests in the current program and from Ernst's tests
is discussed in the following paragraphs. In the effect of variation
in transverse reinforcement on the behavior as by the analysis
developed in Chapter III is discussed.
Be8.ms of the Current Program Without Compression Steel
\ .!'
Figures 51 and 52 show the load deflection curves for beams in the
current program with d = 14 in. and d = 18 in., respectively. Beams J-19 and
J-21 contained #2 U-stirrups rather than the #3 closed rectangular -stirrups of
the companion beams . Figure AA gives the details of the U-stirrupsused in
these 'beams.
. The 'behavior at yield and crushing was not affected significantly by
the change from #3 to #2 stirrups at 6-in. spacing. Since the c-rackpattern
consisted primarily of vertical cracks even at first crushing for these beams,
it is not surprising that the stirrups. had little influeJ:!.ce o.n behavior up to
this point.. beyond crushing, inclined cracks formed
and the s,ize of the stirrups did influence the ultimate deflection.',
The beams w,ith#3 closed stirrups showed considerablY'larger
ultimate deflections thanBeamsJ -19 and ._J -21 with #2, U-stirrups (Figs. 51 and
52).' Beams J-4 and J-9 with d = 18 in. had considerable difference in their
, .". . ," ," . , .; ."
ultimate deflections even though they were practically identical. This is due
in part to lack of symmetry in the failure of J-4 as c?mpared to J-9, as
discussed previously. In view of this observed scatter in ultimate behavior,
only gene:pal trends can be discerned from this very li.rnited of t'ests
with in stirrups .
. #2 -q-stirrups were provided wi t;h rather-large hooks at the top ,
as shown in Fig. A.4, even though they were not closed as were the #3 stirrups.
-83-
-The photographs of Beams J -19 and J -21 after failure in Fig. 53 show the
confinement offered by the hooked ends of the stirrups. The importance of
the position of the stirrups is also illustrated by these photographs. These
stirrups obviously provided some confinement and were located near the point
where crushing was most severe-, Had they been located 6 in. out from the face
of the stub rather than 3 in.: it is doubtful that they would have been as
effective.
For the beams of the current program without compression steel, the
analysis predicts the trend of the observed behavior for the beams in which
the transverse reinforcement varied from #2 U-stirrups to #3 closed stirrups.
The stirrups were not really effective prior to first crushing, and the
analysis is -accordingly unchanged by the variation in-stirrups. At the yield
and crushing stages of behavior, good correlation between computed and measured
values for moments and deflectiops was obtained for these as well as for
the other beams of the current t-est program. Table 11_ shows the comparison of
computed and measured values at yield, and Tables 13 and 19 show the comparison
at crushing.
liUrther tests would be to detennine the comparative merits
of the #2 U-stirrups as compared to #2 closedretangular stirrups. However,
the additional defonnation capacity beyond first crushing, as well as the
appearance of the crushing -pattern of the beams at failure, indicate that
Beams J-19 and J-2l had the benefit of at least some confinement of the concrete
in the compression zone. In predicting the ultimate behavior of these beams
they were analyzed as though they contained closed #2 stirrups. The results
of the analysis forJ-19 and J-2l using this apprqximation compare favorably
with the observed ultimate moments and deflections for these beams as shown
in Tables 21 and 22. The correlation between computed and measured values is
-84-
also satisfactory for. the companion beams with #3 closed Btirrupsas shown
in these same tables.
BeELTI1S of the Current Program with Compres s ion Steel
The load-deflection curves for beams with pl/p = 1.0 and #2 closed
rectangular stirrups are compared in Figs. 51 and 52 with those for companion
beams having #3 closed rectangular stirrups at the same 6-in. spacing. For
these beams, only the size of the stirrQPs varied, rather than both size and
I
shape as was the case for the beams without compression steel.
The behavior of these beams prior to first crushing was not affected
significantly by the change in size of stirrups. Just as in the beams without
compression steel, the crack pattern consisted primarily of vertical cracks up
to this 'point and the stirrups had little or no influence on behavior. As
deflections increased beyond crushing, very prominent inclined cracks formed
and the size of the stirrups did influence the ultimate deflection. In this
case, the influence of transverse reinforcement on ultimate behavior was related
to the II shear strength
ll
contributed by the stirrups rather than the binding
effect which was involved in the failure of bea:.ms without compression steel as
des cribecL above.
'Ilhe beams withp'/p = 1.0 and #2 closed stirrups (J-20 and J-22) and
their companion beams (J -13 and J -6) 'w-ith #,3 closed stirrups all had similar
modes of failure-. In all fourbea.rns, a shearing movement was associated 1-Tith
the final failure, as sholftm in the photographs of Figs. 12 and 13 by the shift
in the original centerline marked on the beams. The smaller measured ultimate
deflections for Beams J-20 and J-22 compared to their companion beams. with
heavier stirrups results from. the lower "shear strength
ll
with lighter stirrups.
These were not classic reinforced concrete shear failures since the full
flexural capacity was d e v e l o p e d ~ and all of the beams were well into the plastic
-85-
range of their load deflection behavior (7 to 13.5 in. deflection) before
failure occurred. The shearing movements associated with these failures
(Figs. 12 and 13) indicate that the stirrups were unable to restrain the
failures in a sense analogous to shear failures. An important difference
which should be noted is that in these beams the crack along which the shearing
movement occurred is more nearly vertical than .is the case for usual shear
failures.
Itis unfortunate that beams failing in this manner were selected
since the effect on the ultimate behavior of the variation in confinement by
the transverse reinforcement cannot be studied. However, the beams with #2
stirrups confirm the mode of failure observed in the companion beams having
#3 stirrups and the trend of reduction in ultimate deformation capacity with
the lighter stirrups is consistent for beams with d = 14 and 18 in.
No attempt has been made to evaluate the shear strength of reinforced
concrete membersin this study. The stirrups provided for the beams of the
current test program were sufficient to prevent usual shear failures. In
providing stirrups to carry all the shear, the calculations usually assume
to be effective those stirrups which are intercepted by a line at 45 deg.
originati:rlg from the critical An analYSis of Beam J-6 with d :::;: 18 in.
and stirrups spaced at 6-in. centers would predict three stirrups to be
effective; if these developed their yield strength,more shear capacity would
be predicted than that observed at ultimate in the test.
An examination of, the photograph of Beam J-6 after failure (Fig. 12a)
indicates that the angle of the critical shear crack with the horizontal is
more than 45 deg.; and, as a result,the shear was not resisted by three stirrups
as predicted. The same behavior is noted from the. photographs of the other
three beams whose failure involved shear (Figs. 12 and 13). Fewer stirrups are
-86-
actually effective in carrying -shear than usua.lly assumed when the failure
crack is inclined at more than 45 deg. to the horizontal as usually assumed
in the truss analogy. Since these stirrups (in some cases only one) cannot
carry the shear force, the beam fails in llshear!Y at ulti.rnate.
These tests would indicate that a closer spacing of stirrups in
the region: of highest curvatures adjacent to the column stub would be more
efficient in preventing this type of failure involving sheaIjwhere the angle
of the. crack along which failure occurs is steeper than 45 deg. Of course,
these more closely spaced closed stirrups would also offer more confinement
to the compression zone in this critical region, making the beam more ductile.
The limiting 'casewouldbereached when the transverse reinforcement is
effective in preventingshearfailures while at the same time provid'iTIg,Y'Sil-f-
ficient confinement to the compression zone to force failure by fracture of
the tension steel. Further tests would bereq,uiredto study fully this inter-
action of shear and m o m e n ~ at ultimate where large plastic deformations are
involved before, failure.
The analysis developed .in Chapter III assumes a flexural failure
by crushing of the concrete and buckling of the compression steel. Beams J ... 20
andJ -22 wi th#2 stirrups, and their 'companion beams, J'-13 and J -6, failed in
the mode involving shear described above and at .consistently smaller deflections
'than-those predicted) as shown in Table 22. It should be noted) however, that
the analysis does predict smaller ultimate deflections for the 'beams 'with #2
stirrups than for t h ~ companion beams wi th #3 stirrups. Beam;j J -,13 showed
less 'effect of shear at failure than the other three 'beams of this group. Its
measured ultimate deflection is 73 peJrcent of the computed value while the
ratios of measured to computed values 'for the others are all smaller than this)
the lowest values being 'for the deeper beams which had higher shear.
-87-
The trends indicated by the analysis are in the right direction,
as indicated above, but the results of the analysis cannot be compared
directly with measured ultimate deflections for these beams because of their
mode of failure.
Ernst's Beams
The amount of transverse reinforcement provided in'Ernst's beams
was varied slightly in each group as shown in Table 3. This was a secondary
variable, the stirrups being:provided to carry the total shear "at a unit
stress well below their yield point." However, the two #3 closed stirrups
at 4 to 6-in. spacing -in the Group 2 beams and the #3 closed.stirrupsat 4.5
to 5-in. spacing 'in the Group 3 beams 'provided cOr,Lsiderable confinement to
the. concrete. None of the beams in the current test program had stirrups which
were as heavy as those provided in these tests. The results of these tests
may be used to check the influen.ce of confinement from transverse reinforcement
assumed in the analysis.
Ernst's tests involved variation in stub length as a major variable.
For the beams of Group 2, three of the members had constant size and spacing
of stirrups while the stub length varied from 6 to 18 in.j for the beams of
Group 3, four of the members had constant size and spacing of stirrups while
the stub length varied from 6 to 24 in. The influence of the length of column
stub may be observed from the behavior of the beams in each group
stirr.ups.:.
Since the span was 9ft for all of Ernst's beams, the distance,,
from support to the face of stub varied with the length of the column stub.
If curvatures in the stub made no contribution to the rotation and thus to the
deflection at the face of the stub, the deflection would decrease as decreased.
The fact that approximately the same ultimate deflection was observed for the
-88-
beams in each group with constant stirrups indicates that the stub did
contribute to the ultimate deflection.
In the ultimate. deflection analysis of these beams, the effect of
the stub vas considered by extending the equivalent curvature distribution a
distance (4.8 + 0 .lw) into the stub. For the beams "W'i th v = 12 in. this
function gives 6 in. = -v.! /2, 'w-hich had been assw!led in the analysis of beams vi th
w =12. in: tested at the University of Illinois. Table 24 shows the c o m ~
parison of' computed and measured ultimate deflections. In general, the analysis
reflects the trend of the o-bserved ultLmate deflections for the beam.s in each
group "iIrb.ich had the same stirrups. The comput.ed values tend to run slightly
higher than the measured ultimate deflections.
The beams which involved the greatest stub length also had more
closely spaced stirrups. The curvature within the stub is assumed to be satis=
factorily t,aken into a.ccount in the analysis by the equivalent di.stri'bution
length (4.8 + 0 .11lT), which gave good resul tsforthe 'beams in each group with
constant st,irrups. Closer stirrl1.p spacing increases the confining effect}
and the analysis predicts a slightly higher ultimate curvature for the section.
Both the additional stlib length and the closer stirrup spacing tend to in-
crease the predicted ul titll.ate deflection. These e:ffects a.re offset by a
reduction in the distance, i.} from support to the stub face} tending to reduce
the ulti:.mate deflection. From the comparison of com:puted and. measured values
given in Table 24, it is noted that the computed ultimate d.ef'lections follov
the trends of the measured ul tLrnate deflections for the beams 'wi th closer
stirrup spacing.
Although the agreement is satsifactorYJ the analysis consistently
slightly overestimates the defor:.mation capacity. This comparison vould tend
to indicate that the confining effect of the stirrups is overestimated
-89-
slightly, although the same error results if the longer column stub actually
contributes less to the ultimate deflection than ,is assumed.
Summary
Any consideration of the effect of transverse reinforcement involves
a large number of variables) only a 'few of which were involved in',the tests
described in this section. The results of the analysis for ultimate deflection
follow the trends observed in the measured values. As more tests are available
the effect of transverse reinforcement may be studied more thoroughly, and
the simplified expressions of Section 14 may be im:proved.
The brief series of tests in the current program described above
emphasizes the importance of the mode of failure on the ultimate deflection.
Further tests may also provide a reliable basis for determining the
and spacing of to insure flexural failure at ultimate,
rather than ,failures involving shear.
23. Effect of Repeated add Reversed Loading
All the beams of the current test program were unloaded at various
points during ,the test to determine the slope of the unloading curve as well
as to see what the load-deflection response of the member would be upon ,re-
loading ,in the same direction. In this section, this process of removing and
reapplying load in the same direction is referred to as II repeatedloading
ll
,
and the average of slopes of the unloading "and reloading curves is referred
to as the slope of the repeated loading
In addition to this repeated loading, three members were subjected
to downward loading 'initially, followed by upward loading. This was
repeated as significant stages of behavior were reached as a result of 'loading
in each direction. Following several cycles of upward and downward loading,
-90-
the member was loaded downward to failure. In this section) this type of
'loadingseCluence is referred to as IIreversed loading) II although repeated
loading in each direction was also applied to these members.
Repeated Loading
The effect of repeated loading was observed from the shape and
,average slope of the unloading and reloading curve. The seCluence varied
slightly;, but the procedure generally followed was to unload shortly after
first crackingoccurred) at about 50 perc-ent of the yield load) inrrnediately
following first yield) and at several points within the plastic range of the
load-deflection history. A continuous plotter record of load versus deflec-
tion was obtained in addition to loads and deflections measured atfre<luent
intervals .wi th the load dym)mometerand dial gages ) respectively) to g-i ve more
accuracy to this record. The continuous record was particularly valuable in
observing the shape of the unloading--reloadingcurves. The average slope
taken from load-deflection records has been used asa measure of the stiff-
ness of the beam at various stages of crushing during the test.
In Section 7 dealing with behavior 'it was pointed out that the
initial slope of the load-deflection curve prior to cracking was a function
of the stiffness of the uncracked section. Figure 54 shows a typical load-
deflection curve to an expanded scale) with the : repeated loading -lines i n d , i ~
cating,the changes in slope as c-racking 'progressed within the elastic -range
of behavior for the beam. The slope of the repeated loading curve showed a
gradual decrease between first cracking and yield. When the beam was unloaded
at approximately 50 percent of the yield load) the slope of the repeated
loading 'curve was less than it had been just following first cracking. There
was a break in theload-'deflection curve upon ,reaching the point to which the
beam had previously been loaded) and at this point the flatter slope of the
-91-
load-deflection envelop.e was resumed just as though the repeated loading had
not occurred. This indicates that the beam gradually changed stiffness as
cracking progressed up to the yield point.
The slopes of the repeated loading curves for points beyond yield
also reflect a gradual reduction in stiffness of the member with increaSing
damage. From this study,the change in slope of the repeated loading curves
was found to be related to the amount of plastic deformation. This .effect is
shown graphically in Fig. 55. The "damage' ratio" plotted on this
figure is taken as the deflection beyond yield prior to unloading, 6:.., divided
l
by the 'plastic deformation ultimately reached in the test, 6:.. This ratio
p
varies from zero at yield to one at ultimate. The "relative stiffness" plotted
vertically is taken as the slope of the repeated loading curve at this point)
e., divided by the slope of the repeated loading curve at yield) e .
l 0
Figure 55 indicates that the "elastic" stiffness decreases almost
linearly to about one-half 'itsinitial value when one-half of the available
plastic ductility has been used up) and remains practically unchanged for
additional deflection. The physical reason for this trend is a gradual in-
crease.in damage due to crushing of theconc-reteand increased cracking as the
plastic deformation increases to about 50 percent of the ultimate capacity. By
this time the concrete cover 'outside the stirrups is completely crushed in the
vicinity of the critical section) and the confined concrete core and compres-
'sion steel are acting together to carry the total compressive force. While
some further damage occurs with additional plastic deformation) this remaining
cross section changes only slightly and the stiffness is therefore almost
constant during this stage of behavior." This empirical relationship should
be useful in .estimatingthestiffnesses and natural periods 'of vibration of
reinforced concrete frames subjected to repeated shocks 'from earthquakes or
blasts.
-92-
It was pointed out above that the repeated loading curve resumed
the shape of the load-deflection envelope upon ,reaching the previous maximum
load level within the elastic range of behavior. Since the beam is relatively
undamaged at this stage) such behavior would be expected. There has been
some question) however) about the capacity of reinforced concrete members to
reach the previous moment capacity upon reloading when ~ 8 e y have been unloaded
after being deformed well beyond yield into a range in which the damage is
very extensive. The current test program indicates very clearly that the
member does possess this capacity for loads up to and even slightly beyond
the maximum. It must be remembered that -all these members were provided with
closed stirrups which gave integrity to the confined concrete core, and that
this behavior would probably not be representative of that ,for beams with no
stirrups or with open stirrups.
Reversed Loading - ,Beam J-3
.Figure 56 shows the load-deflection curve for Beam J-3 which was
'subjected'to reversed loading. Cross-sectional and material properties for
the 'beam are given in Table 1. The loading cycles which were applied are
deSignated as follows:
(a) downward to cracking, upward until crack extended entirely
across the critical section
(b) downward to 50 percent of yield load) upward to 50 percent
of yield load
(c) downward to yield, upward to yield
(d) downward to crushing, (1.6 in.) J upwHrd to crushing (2 in.)
(e) 'downward to deflection of 3-1/2 in., upward to deflection
of 4 in.
() downward to deflection of 9.4 in., upward to deflection
of 9.8 in.
(g) downward to failure.
-93-
The -reduction in-stiffness as a result of previous plastic
deformation in the opposite direction is very obvious from Fig. 56. It should
be noted) however) that the slope of the repeated loading curves for loading
-in one direction only is fairly constant at approximately 50 percent of the
slope following-initial yield. This is particularly noticeable in Cycle (f)
where the maximum deflection was about 10 in. in each direction.
The computed values for moments and deflection at yield, crushing,
and ultimate for Beam J-3 are compared with the measured values in Table 25.
The measured values are influenced by the reversed loading sequence and should
not be considered as very precise quantities. Crushing at the interface of a
crack should not be expected too progress inexactly the same manner as for
concrete which is uncracked. Ultimate behavior is certainly affected by the
reversed loading, but the ability of this member to deform 10 in" in each
direction while continuing to carry increasingly higher moment is remarkable.
Since the direction of -loading was reversed at 10 in. deflection while the
load was still increasing, no "ultimate
ll
was measured in a sense comparable
to the beams loaded in only o n ~ direction. The companion beams (J-8 and J-17)
loaded in only one direction carried their maximum moment at about 16 t-018 in.
deflection.
Failure was a result of buckling of the compression steel when the
beam was loaded downward following the 10 in. upward deflection. At this
stage, the concrete was so severely crushed that the steel was actually
carrying all the moment.
Reversed Loading - -Beam J-7
Figure 57 shows the loading cycles followed in the reversed loading
program for Beam J-7. Cross--sectional and material properties are given in
Table 1. The cycles are designated as follows:
-94-
(a) downward to cracking, upward to cracking
(b) downward to 50 percent of yield, upward to 50
percent of yield
(c) downward to
75
percent of yield, upward to
75
percent of yield
(d) downward to yield, upward to yield
(e) downward to 0.8 in. deflection, upward to 0.8 in. deflection
(f) downward to 0.8 in. deflection, upward to 0.8 in. deflection
(repeat of cycle e)
(g) downward to 3 in
l
, deflection, upward to 3 in. deflection
(Ii) downward to failure.
These cycles differ slightly from the loading sequence followed for
Beam J-3. C,ycles (a), (b), and (c) were all within the elastic range of
behavior, arid cycle (d) was carried to initial yield in each direction. The
plot of load versus deflection for each of these cycles is shown in Fig. 57.
Figure 57 shows cycles (e) and (f) which were carried to
approximately the same deflection (0.8 in.) into the plastic -range in each
direct-ion. The slope of the curve for downward loading is the same for
cycle (e) as for cycle (d), indicating -no loss in stiffness as a result of the
loading to initial yield in the opposite direction. However, the slope of the
upward curve for cycle (e) shows the influence of the O. 8-in. downward deflec-
tion,the slope being less than for cycle Cd). This reduced slope is a result
of the "Bauschingereffect
tl
in the steel (Figs. 58 and 59) ~ s w e l l as reduced
stiffness due to cracking of the concrete. Cycle (f) was a repeat of cycle (e)
to observe the behavior when the same plastic deflection was imposed for a
second time in each direction .. The same trend of reduction in stiffness due
to cracking and Bauschinger effect was noted, the slope of the downward loading
curve for cycle (f) being considerably reduced from that of cycle (e). These
effects also influence the shape of the curve; the curve tends gradually to
-95-
flatten out rather than a sharp yield point. It should be noted)
however) that the unloading curves were almost identical for the two cycles.
The downward loading curve for cycle (g) is identical with that for
cycle (f). This is very important) since it indicates that the reduction in
stiffness observed from the lower slope for cycle (f) downward loading compared
to the slope for cycle (e) is not a continuing effect. Most of the Ildamage
Tf
appears to be done in the first two cycles with a fairly stable load-deflection
response for cycles thereafter. This means that the area under the curve out
to the same deflection, and thus the energy absorbed, would be less for cycle (f)
than cycle (e)) but probably would not change significantly with additional
cycles to the same deflection in each direction.
Cycle (g) was carried to about 3 in. downward deflection) followed
by about 3 in. upward deflection. This was followed by cycle (h) ,which was
downward loading to failure. The slope of the upward loading curve for cycle
(g), following3-in. downward loading, is less than the slope for cycle (f)
upward loading followingO.8-in. downward deflection. The same trend is
observed in the lower slope of the downward loading curve for cycle (h) com-
pared to that for cycles (f) and (g).
The slope of the reloading curves for the same direction of loading
Boes not seem to be influenced as significantly by reversal of loading fol-
lowing plastic deformation as is the slope of the reloading curvein the
opposite direction. This reflects the st.ronginfluence of the Bauschinger
effect" which affects the modulus of elasticity for steel and the shape of the
curve upon loading in the opposite direction) but has only a slight effect on
the modulus and shape of the curve when loading is reapplied in the same
direction. Figures 58 and 59 show measured stress-Btrain curve for S.A.E.1030
steel (8) subjected to reversal of loading. The shape of these curves shows a
-96-
striking resemblance to the load-deflection curves observed in these tests.
They also indicate that the Bauschinger effect is about the same} whether the
first half cycle is in tension or in compression.
Reversed Loading - J-12
Figure 60 shows the load-deflection curves for the reversed loading
seCluence imposed on Beam J-12} and its cross-sectional and material properties
are given in Table 1. The cycles are designated as follows:
(a) downward to cracking) upward to cracking
(b) downward to 50 percent of yield} upward to 50 percent of yield
(c) downward to yield} upward to yield
(d) ,downward to deflection of 2 in., upward to deflection of 2 in.
(e) downward to failure.
This beam was reinforced wi thtwo #6 barS' in the top and two #8 bars in the
bottom in contrast to the sYlli1l1etrical steel arrangement used in Beams J - 3 and
The same trend of loss in stiffness asa result of reversed loading
was observed here as with Beam J-7 which was almost identical except for the
lighter 'top'steel used in J-12. The stiffness reflected by the slope of the
down'ward reloading curve for cycle (d) is only slightly reduced as a result of
loading to initial yield in cycle (c) since there is little damage to the
member at this point. However, the slope of the reloading curve downward for
cycle (e) is greatly reduced by the previous upward loading of 2 in. in
cycle Cd). The reduction in slope upon reloading following 3 in. deflection
in the opposite direction for Beam J-7 was about the same. The Bauschinger
effect accounts for much of this reduction in stiffness, and the greater steel
area contributing to the stiffness ofJ-7 is offset by the larger plastic
-97-
deformation (3 -in. compared to 2 irL) than was imposed on J-12 prior to
reversed loading.
The unequal steel areas produce different moment capac-itiesfor
downward and upward loading 'in the case of J-12, but otherwise itBload-
deflection response follows the same pattern as for the two beams having
symmetrical steel arrangement. The mode of failure for J-12wasby buckling
'ofthe top steel in a vertical plane as in the photograph of Fig . 61a.
This clearlyindic-ates the effect of reversal of loading on the mode of failure
since the companion beam to J-12 which was loaded downward to failure, J-5,
failed by fracture of the tension steel with much less damage to the concrete
(Fig. 61b). It is also interesting to note that the steel buckled upward in
this case rather than outward as for the other beams loaded only downward which
failed 'in this manner (Fig. 11). Buckling upward is opposed to the curvature
of the bars from ,in connection with downward loading and must be a
result 'of '''straightening'' of the steel due to reversed loading. The crushing
of the confined conc-rete 'core exerts a force against the bar which contributes
tb ,:this buckling, as discussed in Section 8.
Comparison of Companion Beams with and Without Reversed Loading
While these three tests represent a very small statistical sample
when we consider the large number of varia'bles which affect the behavior of
reinforced concTetemembers, the trencls which have been observed are quite
significant. The essential requirement of closed ties to support the steel
as well as.to confine the concrete core is emphasized by these tests. When
they are provided,under--reinforced beams possess remarkable ductility.
Figures 62, 63, and 64 show the downward envelope for ,the repeated loading tests
compared with' the load-deflection curves for the companion beams which were
loaded only 'downward to failure. The load levels and ultimate deflections can
-98-
be compared only Clualitatively since the properties of' materials varied somewhat
-and the cycles of reversed loading were different in each case.
Figure 62 shows the comparison of load deflection envelopes for beams
J-3 and J-8 having d = 10 in. The deflection of about 10 in. downward shown
in the figure for Beam J-3 was followed by a 10-in. upward deflection} the
envelope of which was practically identical to the one shown here. This r e p r e ~
'sents ductility in each direction which is about two-thirds of the ductility
for loading to failure downward for this member . Actually) Bleam J --3 was not
loaded to failure at the deflection shown but the loading 'was reversed at this
point. The ductility demonstrated by the member out to this deflection "\;ITould
be' adeCluate for almost any conceivable service reCluirement.
Figure 63 shows the comparison of 'load-deflection envelopes for beams
J -6 and J-7 having d = 18 in. In this case, both beams were loaded downward
to failure, and the two curves have similar shapes. Both beams failed in a
mode. involving -shear, causing'a rapid decrease in load beyond the maximum. The
def"lection at ultimate .load was about 30 percent less for the beam (J "';7): sub-
jected to reversal of 'loading than for the beam (J -6) which was loaded downward
to .failure. The maximum upward deflection imposed on J-7 was 3 in. before
loading downward to failure, but it was 'subjected to several cycles of loading
to deflections less than-3 in. before this as shown in Fig. 57.
F'igure 64 shows the comparison of 'load-deflection envelopes for beams
having d 18 in. but with unsymmetrical steel arrangements. The maximum upward
deflection imposed on Beam J-12 before loading downward to failure was 2 in. as
shown in Fig. 60. In this case, the reversedloading reduced the deflection
at ultimate load by 50 percent and changed the mode of failure from fracture
of the tension steel (J-5) to buckling of the compression steel (J-12) (Fig. 61).
The comparison of these two. envelopes indicates that reversed loading may have
-99-
a more detrimental influence on the ductility of beruus with unsymmetrical
steel (p vip = 0056 in this case) than on beams with syrmnetrical top and bott'om
steel (p sip = 1) 0 HO\vever, since the mode of failure changed) this may not be
a typical case to compare 0 The fai.lure of J'=12 indicates the importance of
the support offered the compression steel by the closed stirrups since buckling
occurred I{i thin a 6-in. stirrup spacing 0
VI. SUMMARY AND CONCLUSIONS
24. Object and Scope
The object of this study was to establish load=deformation relation-
shiI>s for the beam-colunm connections of reinforced concrete frameso The
results of experimental s,tudies together with theoretical considerations have
been utilized in arriving at the methods of analysis presented hereino The
analysis I>redicts four significant points in the load=deformation response of
a beam-column connection: cracking) yield) crushing) and Although
the behavior of a reinforced concrete member with increasing deformation is a
constantly varying phenomenon, these four points mark the significant changes
which are most important in defining the complete response 0
Experimental data from tests of 59 beams in three different test
programs were included in this study 0 Eighteen of these tests were from
McCollisterVs test program (2), 20 from test program (3), and 21 from
the program conducted by the writer. In all three programs, simply supported
beams were loaded through a column stub at midspano Figure I shows the dif-
ferent test specimens used and Figs 0 2 J 3;1 and 4 show the test set-ups 0
The principal variables studied in these three programs
tension steel ratio, p, compression steel ratiO, the ratio pulp, concrete
strength, length of column stub, rate of loading, and loading and
directiono Beams of McCollisterVs and ErnstGs programs were tested on a span
of 9ft while those of the current program had a span of 12 The moment
gradient varied in McCollisterGs tests as a result of variation in the tension
steel percentage, in Ernst
7
s tests as a result of variation in stub length, and
in the current program a of variation in depth.

-101-
The scope of this study , including the results available from three
investigations, is more general than would have been the case for any of the
programs alone. Only under=reinforced beams were included in this study.
25- Behavior
The behavior observed fortheheamsofthe current program, together
with the complete data taken in these tests and McCollister!.s test program,
served asa guide in the formulation of a method of analysis at each of the
four significant stages of behavior. Since the beams were under-reinforced,
a typical seq,uence of these stages of behavior occurred for ,all the beams, as
-shown in the generalized load-deflection curve of Fig. 5.
The ,first break in the load-deflection curve occurs when the beam
c-racks) as represented by . point 1 on the curve of Fig . 5. Beyond this point,
the stiffness of the beam changes with 'cracking 'from that of the uncracked
section to that oftheeracked section, and the slope of the load-deflection
curve changes accordingly between points land 2.
The second significant stage is the yield point of the member,
represented by. point '2. Since only under-reinforced members are considered
in this study, the tension steel yields before the crushing strain is reached
in the concrete. The yield point rnarksthe boundary between elastic and
inelastic behavior as observed from the load-deflection envelope for the beam.
With intermediate grade steel) there is very little increase in load as
deflec'tionsincrease 'beyond the yield point until the tension steel strain
enters the strain-hardening -range and the def'lection becomes considerably
larger than the yield value.
Point 3 on the curve of Fig. 5 represents first crushing 'of the
concrete. Theload ... deflection curve cannot be used to detect crushing since
there is no significant change in the curve at this point. The visual

"appearance of crushing "is that of surface spalling :on the t-op of the beam
adjacent to the stub) which gradually increases in extent as deflections
increase. Strain gage readings on the concrete , "as well as visual appearance,
were used to detect crushing as defined in this study. The gages mounted on the
top face of the beam one inch out from the stub face showed increasingcompres"
s"i ve st-rainsuntil first crushing "occurred" at which stage the readings began
t-o reverse. Plots of deflection versus concrete strain such as those given in
Fig. 7 show this reversal rather "clearly. Figure 8 shows photographs of the
visual appearance of the crushing -as defined herein.
The ultimate load capacity for the beam is the final behaviorst-age
"
cons ideredin this study; point 40f "Fig. I'represents "this stage . There is a
rapid decrease in load 't,yi thincrease in deflection beyond this point" except
for some 'beams without compression steel for whic,h failure was somewhat more
gradual. Different modes of failure 1ATere a-bserved crushing of the confined
concrete core for beams wi compression steel (Fig. 9)" crushing of the
confined concrete and buckling "of compression steel (Fig. 11), and buckling of
compression steel combined with a shearing movement (Fig. 12). These modes of
failure at ultimate are discussed in Section 8.
26. Analysis and Resu:lts
Procedures for computing moments and deformations at each of the
four significant stages are developed and discussed in Chapter III; Comparisons
of the computed and measured values are presented in Chapter IV.
Cracking Stage
The beam is considered t-o behave elastically as an uncracked cross
section prior to cracki"ng, and the distribution of curvature along -the span
follow's the moment diagram as shown in Fig. 24a. The cracking "load is reached
-103-
when the computed extreme fiber stress at the critical section is equal to
the modulus of rupture of the concrete. Thus
where the modulus of rupture" f , is assumed to be related to the concrete
r
strength as follows
f
r
3
+'12,,000
ff
c
(2)
The relationship between the load and deflection prior ,to cracking
'follows ,from the curvature distribution of Fig. 24a. The deflection at first
cracking cis then
where I
E
c
gross moment of inertia of the cross section
30)000
10
6 + fT
c
The computed loa,d and deflection at cracking are shown on the
(1)
expanded early portion of the measured load-deflection curves in Appendix C.
The agreement with the measured break in the curves is satisfactory) and a
better approximation of the load-deflection response of the member maybe made
by considering this point rather than neglecting it and assuming linear elastic
behavior up to yield" as is often done.
Yield Stage
This stage represents the yield point in the observed load-deflection
response for the member which results from yielding of the tension steel. In
the analysis" the ordinary straight line theory was used to locate the position
of 'the neutral axis (Eq. 4)" assuming the modular ratio" n) to be a function of
-104-
concrete strength as follows
10
n 6 + fr
c
The yield moment is
M = CAf - A if I) Cd - :kd) + A I fJ (d-.d I)
Y sy ss 3 ss
Figure 35 shows a plot of measured versus computed yieldrrioments for beams
tested at the University of 'Illinois; the agreement is within + 5 perc-ent for
all but a few cases.
Once the ,neutral axis is located} the curvature at yield" cp. " may
y
be .found as
where
E
sy
CPy
E
sy
d - kd
fy
~ ~ ~ + 0.0003
30,000
(6)
(8)
This yield strain is different from that usually used, and the 0.0003 increment
may be thought of as correcting for the differences between real and assumed
stress--strain relationships for the intermediate grade steel bars as well as
fortheerror 'in computing :theposition .ofthe neutral axis by the straight
'line t h e o r y ~ These factors 'are discussed in Section 12.
The distribution of curvature at yield is assumed as shown in
Fig. 24"0, and the yield deflec-tion is thus
where w = length of the 'column stub.
Figure 36 shows graphically the comparison of measured and computed yield
deflections. The agreement is within + 10 percent for all but two of the beams
tested at the University of 'Illinois.
-105-
The above ana.lysis -would not ~ b e valid for .'beams with high values of q.
In this case, .theneutral axis must be located by assuming a realistic stress-
strain curve for the concrete rather than the .linear relationship. or straight-
line theory.
Crushing Stage
The l:i:m.i ting strain in the concrete associated wi th crushingin this
study was 0.004. The intermediate grade steel was assumed to follow the
complete generalized stress--strain curve of Fig. 27,a.nd the average concrete
stress at erushing "",TaS assumed as
f
cu
where f! andf are both in psi.
c cu
(10)
The point of action of the compressive force is assumed to be 0.42 kd
c
from the top of the be8J!1. Strain at the level of the tension steel is assumed
to be FE) whereE is the steel strain fora .linear distribution of strains
s s
over the . section. In this analys is ) ~ .. tb..e reduction factor , F) [Nas taken as 0.85.
'llhe neutral axis isloc.atedby assuming trial values of k d until the
c
forces acting on the section are ineq,uilibrium. The concrete is assumed to
carry no tension.
The crushing moment is
M
c
(A f
. ss
Figure 37 shows the plot of measured versus computed moments at crushing
utilizing this analysiS, the agreement being within ~ 5 percent.
Having located the neutral axis) the curvature at crushing is
0.004
k d
c
(12)
(11)
-106-
The deflection atcrushingwascomputed using three differentequi valent
distributions of curvature. Methods 1 and 2 account for the additional
deflection beyond yield) 6 - 6 ) which is then combined with the computed 6
c y y
to determine the crushing deflection) 6. Method 3 is a very approximate
c
method of predicting the total deflection at crushing.
Method 1
The distribution length for peak curvature into the s'P9.-u) 0:) is
assumed to be related to the distribution of moment as shown in Fig. 28. The
in excess of the yield deflection is
where ex
(6- 6 )
c y
i (1 - M 1M ).
y c
K i (cp _.cp) (.::'2
i
[ + a)
cc y , .
K shape factor = (0.9 . ..;.. o. :> 00 5
c
(14)
Figure 29 shows the plot of Kc versus ex where measured values of (6 -6)
c y
were used inEq. 13 to evaluateK .
c
The crushing deflection is found by combining (6 - 6 ) from Eq. 13
. c y
with 6 from Eq. 9. Figure 38 shows the plot ofrneasured versus computed
y
crushing def'lectionsfrom the analysis of Method 1. ' .. ;';is'w:tthin
about + 10 percent.
Method 2
where
The deflection in excess of the yield deflection is
(6 - ) = xii(cp-'cp )
c y c y
spread length of equivalent rectangular distribution as shown
inFig30a .
. Figure 31 shows the plot of computed XV using measured (6 -6) in
c y
Eq. 15 versus tension steel .p) which was used to evaluate Xi for this
analysis 0 This analysis approximates 'x I as follows
-,107-
x! = (10 - ,1. 75p) 5' 6.5 in. (16)
where p is the tension steel ratio in ,percent.
The deflection beyond yield, (Dc -By)) is combined with B from
y
Eq. 9 to obtain the crushing deflection. Figure 39sholfiTs the plot of measured
versus computed crus'hing deflections utilizing Method 2. The agreement is
within .. 10 percent gene rally, just as with Method 1. However, this method is
perhaps more empirical than Method 1.
Method 3
This method is a simplification of Method 2 as sho"(hffi in Fig. 30b. An
equivalent'rectangular distri'bution of the crushing 'curvature, cp , in the
c
vicinity of the column stub accounts directly for the m.easured crushing'deflec-
tion lfiTi thout involving a separatecalculat ion of D. The additional distri'bu-
y
tion 'length, 0 i' which must be added to the length Xl from Method 2 is plotted
x
versusp in Fig. 32. The trend of these ya,lues is
o! 175p < 35
x
When this is combined lfiTi th the trend ofx i values from Eq. ,16, a constant
spread 'length of 'lO,:i;nQis obtained for Method 3. Thus,
where x
D
c
x. cp
c
10 in. = spread length of eg"ui valent rectangular distribution.
Figure 40 shmfS the plot of mea,8ured versus computed crushing
deflections using Method 3; the agreement is w,'ithin + 15 percent .
Ulti.rnate Stage
(17)
This analysis was developed for bea:.ras having closed stirrups or
ties which act as binders to confine a concrete core at ultimate. The average
concrete stress at ultimate, f J and the strain at ultimate at the level of
eU
-108-
the compression steel, E! , are expressed in terms of a binding ratio which
su
is defined as
volume of stirrup + 0.1 (\.Dr/s)'
Pb = volume of bound concrete
where
Dr
diameter of compression steel
s spacing of closed stirrups
In terms of this. parameter
and 0.011 + 0.2 P
b
The steel is assumed to follow t.he complete generalized stress -strain
curve of Fig. 27 except that strain hardening of the compression steel is
neglected. The distribution of strains across the section is assumed linear,
and. only the confined concrete within the stirrup is considered to be effective
in carrying compression. The position of the neutral axis is located by taking
trial values of k d until the tension and compression forces acting on the
u
section are in eg:ui1ibrium.
or
where
The ultimate moment is
M
u
C' (d-d') + C
s c
D'
M = A If! .1\"' d - d I) + ('A f - A! f r ) (d - d! + - o. 45a )
u s s s s s s' , 2
a = depth to neutral axis from inside of closed stirrup.
(20)
Figure 41.shows the plot of measured versus computed ultimate moments, all of
the values agreeing within . 10 percent.
The ultimate deflection analysis generally follows the approach of
Method 1 used for analysis of crushing deflection. In this case, the spread
length of peak curvatures into the span, f3, is a function of moment as f o l l o w s ~
-109-
f3
= (1 -M /M . )
. c U'
(21)
However) the total deflection and curvature at ultimate are computed directly
as 'I;ITasdonein Method 3 for crushing deflection . Ilhustheform of the
expression is
6 = i. lcp ( ~ + A)
U u u'2 !-'
(22)
Figure 34 showsthecomputedK .. values from' Eg,. 22 using :measured ultimate
u.
deflections plotted versusp/!p. The trend of these points g-ives the expres-
sian for the shape factor K used in the analysis
u
K
u
, p!" ~ -
= (15 - .07 -) > 0.8
p
Figure 42 'shows the plot of measured versus computed ultimate
deflections. The opencirc-les which indicate points falling 'outside the + 20
percent 'band on the curve represent beams failing in a mode which involved
shear as discussed in Section 22.
27. Effect of Va.riables
The effects of depth, compression steel) transverse reinforcement,
and repeated and reversed loading are discussed in Chapter V. These were
variables.in the current program) and the comparison of measured load-deflection.
curves for companion beams was used to shov.theinfluence of each of these
variables. O:nlylimi ted data from tests involving variation in transverse
re.inforcement were available) but this is presented along wi tht-rendspre-
dicted in the analysis in Section 22. Only three beams were subjected to
reversed 'loading, and these tests are discussed individually in Section 23.
The other variables were more thoroughly studied in this program.
There is a:happarent interrelationship b e b ~ T e e n the influence on
ultimate deflection of depth and the amount of compression steel (pi/p ), as
-110-
beams with prjp = 0 and 0.56, but for beams Withp'/p = 1.0 the ultimate
deflection d_ecreases with an increase in depth. The ultimate deflection and
the ultimate moment for these 'beams with constant area of tension steel tends
to increase with an increase in depth or amount of compression steel, but there
appears to be a li.mi tto this trend in each case.
The transverse reinforcement may influence the ultimate deflection
in two different vrays. The closed stirrups confine a concrete core which is
effective .in resisting compreSSion at:,ultimate:,af'ter ',crushing
is very severe. Thus the size and spacing of these stirrups is important in
connection vrith their function as binders. At the same time) the stirrups
influence the shear strength of the member. In the current program) some
members failed with a mode of failure involving shear) and the ultimat,e deflec-
tion was less than that predicted by the analysis 'which was developed on the
basis of flexural failures. These failures were not classic II shearfailures
H
since they occurred after several inches of deflection in the plastic -range.
These effects of transverse reinforcement are discussed in Section 22.
All of the 'beams in the current program were subjected to repeated
loading) the load being -removed andre-applied in the same direction at several
points during the test. llhis 'was d,one 'both prior to and following yield, and
the change in stiffness of the member 'at various stages was observed from the
slope of these unloading-reloading curves. Three members were subjected to
both repeated and .. reversed loadin'gto observe their behavior. These beams are
compared with companion beams loaded in only one direction by comparing the
load-deflection envelopes. The strength is not reduced but there is a loss in
ductility as a result of reversed loading ,for 'all three of these beams) as
discussed in Section 23.
-111=
28. Conclusions
The concept of first crushing as the ultimate stage in the load-
deformation response of an under-reinforced concrete member neglects a large
reverse in ductility between crushing and ultimate. This is especially true
for beams with compression steel for which the ultimate deflection may be 10
or 15 times as large as the deflection at first crushing. The ulti.rnate moment
is also greater than the crushing moment as a result of strain hardening of
the tension steel.
Even beams with no compression steel but with low values of q possess
considerable deformation capacity beyond the point of first crushing) the
deflection at ultimate load being two to five times that at crushing. The load-
deflection curves themselves show no indication of first crushing for the
merribers analyzed in this study . This stage is defined .. empirically and repre-
sentsvery minor damage) perhaps no more serious than the tension cracking at
the same load. As more reliable methods are availa'ble to predict the ultimate
behavior of reinforced concrete members) we will be able to compute more
realistically their load-deflection behavior. The point of first crushing
viII no doubt allfiTaysbe of interest in analys is) but it should, not be considered
as 'ultimate in the usual sense of the vord..
The principal factors which contribute to increasing the ductility
of under-reinforced concrete members are compression reinforcement and closed
stirrups or binders which serve to confine a concrete core at ultimate and also
to restrain the compression steel. Perhaps the surest way of obtaining ductility
is to provide compression steel together vith closed stirrups, both of which
contribute to the ultimate deformation capacity. The compression steel itself
confines the concrete within the stirrups, and the stirrups restrain the
compression steel against buckling.
-112- -
The spread of peak curvature with deflection beyond yield is
concentrated in the immediate vicinity of the column stub as observed from the
measured distributions of curvature in the currentprogra.11. The spread of this
curvature into the span vas related to the moment gradient in the analYSis at
crushing and at ultimate, w.ith good results. It is important to note that
large curvatureS1lTere measured 'within the stub in every case. was con-
sid,ered in the analysis at 'both crushing and ult:trnate 'by determining -empirically
a shape factor for the distribution of curvature wi thin an enclosing Tectangle _,
1Irith ordinate'_ of computed curvature, cp or -cp ,and a length extending ..from the
c u
of the stub to the point of theoretical sprea:d into the span, ex
or 13
Tests of frames . may show thci,t less curvature is
present at a -joint than in the column stub of these beams.. This could be taken
into account 'by reducing the assumed spread length, the most conservative
estLmate being to consider only the length ex or -13. The writer feels that this-
might be too conservative but it is one bound for -the ansverin analysis. Loss
of bond along the tension steel contributes to the rotation at the joint so
that at least some equivalent spread length into the stub or -column should be
considered.
Beams having a constant moment over a considerable length should not
be eX,pected to develop curvatures as large as CPu .throughout the region of
constant moment . For -the bea.m-colunLn connection, the critical section is very
localized and the plastic deformation must occur -in this vicinity. On the
other hand, a beam with a-region of constant moment may beg-in to fail at one
section wi thin the region and'continue to fail at that section without much
further damage to the rest of the member . IJ:Jhere is a reserve beyond crushing,
as observed by Gaston (1), but oit is not as large . as that for the beam.=column
connections studied herein.
-113-
Further tests on the conf'iningerfect of closed rectangular 'Stirrups
as well as compression steel in conjunction with them should enable us to
refine the ultimate analysis which is presented in this study. However) the
use which is to be made of the ultimate stage prediction is important in con-
Bidering the accuracy desired. If only an estimate of the reserve deflection
beyond first crushingis desired) the + 20 percent accuracy obtained herein in
predicting ultimate deflection is adequate if it is proved by other tests to
be generally applicable. From the S'tandpointof energy absorbed by the member)
or toughness) the error 'is aLrnostdirectly related to the error-in predicting
the ultimate deflection.
The ultimate deformation capacity of a properly detailed under-
reinforced concrete member -is 'more than ade'luate for theusual rotation required
to provide redistri'bution of moment assumed in limit analysis for usual frames.
REFERENCES
1. Gaston, J. R., C. P. Siess, and N. M.New..rnark, llAn Investigation of the
Load-Deformation Characteristics of Reinforced Concrete Beams up to the
Point of Failure,!! Civil Engineering Studies, Structural Research Series
No. 40, University of Illinois, Urbana, Illinois, December 1952 ..
2. McCollister, H. M., C. P. Siess) and N. M. Newmark, "Load-Deformation
Characteristics of Simulated Beam-Column Connections in Reinforced
Concrete," Civil Engineering -Studies, Structural Research Series No. 76,
University of Illinois, Urbana, Illinois, June 1954.'
3. Ernst, G. C., "Plastic Hinging at the Intersection of Beams and Columns,TT
ACI Journal) V. 28, No. 12, June 1957, (Proceedings V. 53)) pp. 1119-1144.
4. Warwaruk, J.,M.A. Sozen, andC. P. Siess, "Strength and Behavior'in
Flexure of Prestressed Concrete Beams," Civil Engineering Studies,
St'ructural Research Series No. 205, Uni vers i ty of Illinois , Urbana, Illinois,
September 1960.
5. Hognestad, E., N. W. Hanson, andD. McHenry, "Concrete Stress Distribution
in Ultimate Strength De Ejign, V. 27, No.4, December 1955,
(Proceedings V. 52), pp. 455-479.
6. Swihart, G.R., J. R. Allgood, and 'W'. A. Shaw, "Elastic Resistance of
Reinforced Concrete Beams,": Proceedings of the American Society of Civil
Engineers, Vol; 85, No. ST1, January 1959.
7. Chan, W. W. L., "The Ultimate Strength and Deformation of Plastic Hinges
in Reinforced Concrete Frameworks,lt Magazine of Concrete Research, V. 7,
No. 21, November 1955.
8. Dubuc, J., "Plastic Fatigue Under Cyclic Stress and Cyclic Strain With a
Study. of the Bauschinger Effect," Ph.D. Thesis) Department of Strength
of Materials, University of Montreal, 1961.
-114-
-115-
TABLE 1
PROPERTIES OF BEAMS FROM THE CURRENT TEST PROGRAM
(b= 8 in. for all beams except as noted)
C-ylinder Reinforcement Yield Point of
Beam Strength Quantitl and Size Reinforcement {ksi) d
d!
Stirrup Size
Number
f!
(psi) Tens. Camp. Tens. Camp. in. in. and Spacing
c
J-l 4930 2-#8 476
10.0 2.0 #3 at
6 in,
3-11 4110 2-#8 469
10.0 2.0 #3 at 6 in.
3-2 4080 2-#8 2-:#6 48.0 48.6 10.0 2.0 #5 at 6 in.
3-3 4900 2 ... #8 2-#8 48.3 48.0 10.0 2.0 #3 at 6 in.
J-8 4680 2-#8
2.,#8
454 455
10.0 2.0
#3 at 6 in.
3-17 3900 2-#8 2-#8 469
46.8 10.0 2.0 #3 at 6 in.
J-18* 4410 2..,#8 2-#8 454 471
10.0 2.0
#3 at 6 in.
J-24* 5000 2-#4
2..,#4
485 478
10.0 2.0 #3 at 6 in.
J-10
3590 2-#8 451
14.0 2.0 #3 at 6 in.
J-14 4500 2-#6 471
50.0 14.0 2.0 #3 at 6 in.
J-13 4800 2-#8 2-#8 456
46.0 14.0 2.0 #3 at 6 in.
3-19 3900 2-#8 458
14.0 2.0 #2U at 6 in.
J-20 4380 2-#8 2-#8 458 465
14.0 2.0 #2 at 6 in.
J'-4
, ___ ," __ .. __... ___ g.:.#.8--.. -.. .,. ___---.:: . .:::.-----.....---J-.i .. ___
18.0 2.0 #3 at 6 in.
---_.---_ .. _-.. -
J-9 4190 2-#8 470
18.0 2.0 #3 at '6 in.
J' C'
-) 5000 2-#8 2-#6 451 489
18.0 2.0
#3 at 6 in.
3-12 4550 2-#8 2-#6 451 497
'18.0 2.0 #3 at 6 in.
3-6 5160 2-#8 2-#8 46.2 46.4 18.0 2.0 #3 at 6 in.
" f'"7
u-( 4450 2-#8 2-#8 4605 463
18.0 2.0
#3 at 6 in.
3-21 4350 2-.#8 476
18.0 2.0 #2U at 6 in.
3-22 4420 2-#8 2-#8 46.2 46.4 18.0 2.0 #2 at 6 in.
'*b = 6 in.
-116-
TABLE 2
PROPERTIES OF BEAMS . FROM McCOLLISTERiS TEST PROGRAM
(b = 6 in. for all be&"TI.s)
Cylinder Reinforcement Yield Point of
Beam Strength Quantitl and Size Reinforcement

d.
.d
l
. .stirrupSize
Number
fl
(psi) Tens. Camp. Tens. Camp. in. in. and .Spacing
c
8-6 4151 3-#4
44.8 1072
.....
--.--
8-7 4073 2-#4 450 1072 -----------
8-8 2642
450 1072
-_.-.- __ ...... _- ... -
T-l 3897 2-#6 2-#4 41.8
479 1058 1.28 #3 at 6 in.
T-2 3858 2-#7 454
50.0
1037
1.41 #'3 at 4 in.
T-3
4266 2-#11 2-.#8
447
46.1 10.20
156 #3 at 3 in.
T-7 4540 2-#6 4Q9
48.5
1
#3 at 6 in.
T-I0 4330 2-#6 42.6 46.3 1058 133 #3 at 6 in.
T--ll 4470 2-#6 2-#10 425
46.1 1058 1.84 #3 at 6 in.
T-12 4367
2-#9
46.0
458
10.28 1.84 #3 at 4 in.
T-13 4847 2"'#9
2..,#10
569
46.0
1037 1.69 #3 at 4 in.
T-14 4030 2..,#6 41.4
409 1058 152 #3 at 6 in.
T-15 3700 2-#6 2-#6 470 405 1058 150 #3 at 6 in.
T-4 2230 2-#5 2-#4 471 457
10.65 1.28
#3 at 6in.
T-5
2021
2-#7 2-#.5 48.4 46.6
1051 135 #3 at 6 in.
T-6 1905 2-#9 2-#7 565 497 1037 1.49 #.3 at 4 in.
T-8 2440 2-#6 2-#7 450 492 1058 174 #3 at 6 in.
T-9 2693 2 .. #4 2-#6 525 415 1072 154 #.3 at 6 in.
-117-
TABLE 3
PROPERTIES OF BEAMS 'FROM ERNST' STEST PROGRAM
(b = 6 in. for all beams)
Cylinder Reinforcement Yield Point of
Beam Strength Quantity and Size . Reinforcement (ksi) et. d' Stirrup Size
Number
f!
(psi) Tens. Camp. : :Tens. ,Comp. in. in. and Spacing
c
6S-2 4170
3-#7 2-#3 450 455
10.0
15 2-#3 at. 6 in.
12S-2 4170
3-#7 2-#3 450 455
10.0
15
at 6 in.
18s-2 4170
3-#7 2-#3 450 455
10.0
15 2-#3 at 6 in.
248-2 4170
3-#7 2-#3 450 455
10.0
15 2-:<#3 at5 in.
368-2 4170
3-#7 2-#3 450 455
10.0
15 2-#3 at
h .
. In.
6F-2 4710
3-#7 2-#3 450 45::5
10.0
15 2-#3 at 6 in.
12F-2 4710
3-#7
2 ... #3 450
10.0
15
2...,#3 at 6 in.
18F-2 4710
3-#7 2-#3 450 455
10.0
15 2-#3 at 6 in.
24F-2 4710
3-#7 2-#3 450 455
10.0
15 2-#3 at
5
in.
36F-2 4710
3-#7 2-#3 450 455
10.0
15
2...,#3 at 4 in.
3300 2 ... #3 455
10.0
15 1-#3 at 5 in.
12.8-3 3300 2-#5 $,1#3 475 495
10.0
15 1-#3 at ,5 in.
18S-3 3300 2-#5 2-#3 475 455
10.0
15 1-#3 at5 in.
24s-3 3300 2 ... #5 2-#3 475 455
10.0
15 1-#3 at 5 in.
368-3 3300 2-#5 2-#3 475 455
10.0
15 1-#3 at 4.5in.
6F-3 '3230 2-#5 2-#3 475 455
10.0
15 1-#3 at5 in.
12F-3 3230 .2-#5 2-#3 475 455
10.0
15 1-#3 at5 in.
18F-3 3230 '2-#5 2 ... #3
475 455
10.0
15 1-#3 at 5 in.
24F-3 3230 2-#5 2-#3 475 455
10.0
15 1-#3 at5 in.
36F-3 3230 2-#5 2-#3 475 455
10.0
15 1-#3 at 4.5 in.
-118-
TABLE 4
RANGE OF VARIABLES STUDIED INTIrE TESTPROGRAlV1S
Variable
; Range Studied
McCollister
Y
s Tests ErnstisTests Current Tests
Tension Steel
0.62-5.10% 1.03-500% 1.10-1.98%
Ratio) p
Compression Steel
0-4.00%
Constant
O ~ 1 9 8 %
Ration)
pi
037%
Concrete Cylinder
1910-4850 psi 3230-4710 psi
3590-5160
psi
Strength)
f!
c
Ratio pl/p 0-2.88 0.074-0359 0-1.0
pi' -
p'f!
q'
=
Y
y
-0.280-0.449 0.098-0.489
. 0-0.226
fl
C
Depth, d 10.20-1072 in.
Constant
10.0-18.0 in.
10 in.
Width) b
Constant Constant
6 and 8 in.
6 in. 6 in.
Length of Column Constant
6.0-36.0 in.
Constant
Stub) 'W' 12 in. 12 in.
Span
Constant Constant Constant
9
ft.
9
ft. 12ft.
Loading Sequence, Some Beams Load .in one R.epeated and
and Direction Loaded in the Direction to Reversed
Reversed Direction Failure Loading
Rate of Slo1-T Slow Slow'
*.
Loading (Several Hours) (Servera1 Hours) (1 to
3
days)
Fast
(3 to 6.minutes)
*
"Tests extended over this period, but the load was removed at intervals
during the test and overnight for tests extending over more than one
day.
-119-
TABLE '5
COMPUTED K AND x
1
VALUES 'FOR BEA..TvISOF THE CURRENT TEST PROGRAlVI
c
[from Eqs. 13 and 15 using measured values of (6 -6 )]
c y
ComE' (in.-k)
*
( ~ +0)
Comput.ed (1/in. )xl0=3
Measured Compo Compo
Beam a
6 -6
XI
No.
M M
(epe-cry)
c y
K
Y
c (in. ) (in. )
cry
epe
(in. ) c (in. )
J-l 637 699 59 119 033 156 1.23 051 05,3 63
J-ll 634 656.
23 83 033
1.47 1 ~ 1 4 0.44
070
59
J-2 644 698
52
11.2
032 170 138 053 052 58
3-8 607 647 4.0 10.0 0.29 190
1.61
073
0.69
69
J'-.17 626 662
37 97 031
1.82
151 077
0.80
77
J'-18
599
628
30 90 031 175
.1.44
074
0.86
78
J-24 174 '201
89 149
0.26
258 232 130 057 85
J-10 869 913 33 93
0.21
139
.1.18 0.64 0.88'
8.2
J-14 913 978
4.4 10.4 0.21
179 158 070 0.65
67
J-13 885 1004
79 169
0.20 1.86 1.66
092 050
8.4
J'-19 883 932 36 96
0.21 1.44 1.23 054 0.69
67
J'-20 888 " 969
55 115
0.20 1.85 1.65 0.84 0.67
77
J-4 1132 1274
73 133
0.15 1.60 1.4-5 0.62 0.49 6.5
J-9
1182 1304 6.2 12.2 0.16 1.42 1.26 0.69 0.68
8. ~ 3
J-5 1145 1368 107 167 0.15 178 1.63 0.88 0.49 8.2
J",",6
1177 1401 10.6 16.6 0.15 1.84 1.69 096 0.,52 8.6
J-21 1198 1301
53
' 1103 0.16 1.46
130 0.67 0.69
78
J'-22 1176 1413 11.1
171
0.15
179
1.64
097 053 89
*
M
a =
i (1 _J:..)
M
c
-120-
TABLE 6
COMPUTED K
AND Xl VALUES ':FORBEAMS OF McCOLLISTER j S . TEST PROGRKM
c
[from Eg.s. 13 and 15 using measured values of (6 - 6
y
) J
c
Camp. (in. -k)
*
(w '
Computed (1/ in. )xl0-3 Measured. Camp. Camp.
Beam ex
2'+ex)
6

No.
M
M (qJc
c
Y
K
Xl
c (in. ) (in. )
cry ere
(in. ) c (in. )
8-6 257 320 94 154
0.25 258 23.3 095 055
'85
8-7 175 236 125 185 0.2.3
351 328 157 054
10.0
8-8 174
220
10 .. ;1"
16.1 0.24
274 250 098 051
;08.2
'11-1 342 409
79 139
0.25 2.47 2.22
094
0.64
8
T-2
801
833 19 '(79 032 178
1.46 0.48 0.87
69
T-3 1190 1221
13 73 037
1.25 0.88 0.27 0.87 6.4
T-7 335 392 70 130
0.24 2.63
239 093
0.62 8.1
T-I0 348 403 65 125
0.26
237
2.11 0.82 0.65 8 .. 1
T-l1 338 404
79 139
0.23 2.16
193
0.84 0.65
90
T-12 1000 1032 15 75
034 159
1.25 036
0.80 6.0
T-13
1000 1054
37 97
036
1.82 1.46 0.46 0.68 6.6
TI-14
338 374
4.6 10.6 0.24 2.48 2.24
075
0.66
70
T-15 383 431
55 115 0.27
235
2.08 0.81
070
8.1
T-4 276 329
77 137
0.26
237
2.11 0.87 0.63 8.6
T-5 526 549 19 79
032
1.61 1.29 036 073
,58
T-8 .361 411
58
11.8 0.26 2.04- 178 071 070 83
T-9
201
257 105 165
0.26
255
2.29
093 051 8.4

-M
ex
=
(1
-
i)
. c
-121-
TABLE
7
EQUIVALENT SPREAD LENGTH FOR CRUSHING DEFLECTION ANALYSIS
BY METHOD 3 -BEAMS OF THE crLIRRENT TEST PROGRAM
*
Beam
p(%)
x 10-
3
x 10-:3
qJy/CPc
Method 2
Xl
Method 3
No.
qJy qJc
Xi (in.) (in. ) x ~ i n . )
J-1 198 0326 156
0.209
63
4.6
109
J-l1 198 0330 ~ . 4 7
0.224
59 50 109
J-2 198 0316 170
0.186
58
4.2 10.0
J-8 198 0.294 190 0.155 69 33
10.2
J-17 198 0305
1.82 0.168
77 35
11.2
J-18 2.64 0311 175
0.178
78 3,6 11.4
J-24 0.67 0.263 258
0.102
85
2.0
105
J'-10 1.41 0.211
139
0.152 8.2
31 113
J-14 1.41 0.207
179
0.116
67 25 92
J-13 1.41 0.196 1.86 0.105 8.4 2.1
105
:r _l q
1,41 0,212 1.44 0.147
67 32 99
~ -.,,-
J-20 1.41 0.197 1.85 0.107
77
2.2
99
J-4 1.10 0.152 1.60 0.095 65
2.1 ''8.6
J-9
1.10 0.159 1.42 0.112
83
2.2
105
J'-5 1.10 0.147 178 0.083 8.2
17 99
J...,6 1.10 0.147 1.84 0.080 8.6 1.6 10.2
J-21 1.10 0.160 1.46 0.110
78 2,3
10.1
J-,22 1.10 0.148
179
0.083
89
1.6
105
*
=3c
Mean
103
5 (283 - -x')
Xl
qJc
8.6
Range
11.4
-122-
8
EQUIVALENT SPREAD LENGTH FOR CRUSHING DEFLECTION . ANALYSIS
BYMETHOD 3 -BEAMS OF MeCOLLISTER'S TEST PROGRAM
* Beam
p(%)
x 10-
3
x 10-
3
CPy/CPe
Method 2
x
Y Method 3
No.
CPy CPe
Xi (in.) (in. ) x )
8-6
093
0,248
258 0.096
8 r
.J 13 98
8-7
0.62 0.232
351
0,066 10.0 0.8 10.8
8 ... 8 0.62 0.238
274 0.087 8.2 1.2
94
T-1 139
0.250 2.47 0.101 8.8 1.4 10.2
T-2. 321 0322 178
0.181
69
2.8
97
T-'3 510 0365 1.25 0.292 6.4
47
11.1
T-7 139
0.236 2.63 0.090 8.1
13 94
T-10
139
0.257 237
0.108 8.1 1.6
97
T-11 139
0.231 2.16 0.107 10.1
13
11.4
T-12 4.11
0338 159
0.213 6.0
35 95
T-13 321 0358
1.82 0.197 6.6
31 97
T-14 139
0.240 2.48 0.097 70 15 85
T-15 139
0.268
235
0.114 8.1 1.6
97
T-4 097
0.263 .2,37 0.111 8.6
15
10.1
190 0317
1.61 0.197 58 33 91
T-8
139
0.260 2.04 0.127
83
1.8 10.1
T-9
0.62 0.256
255
0.100 8.4 1.4
98
*
cp Mean
.9"9

Xi
= ...L (22 . 4 - . X I )
CPe
Range
85
11.4
, .-123-
TABLE
9
. COMHJTED 'K VALUES . FOR ULTIMATE DEFLECTION ANALYSIS
-
u
BEAMS OF THE CURR.ENT TEST PROGFAM
Comp. (in. -kips)
*
( + (3)
.
Meas.
**
Beam M 1M
f3
1 K
No. M M
c u
(in. )
cp, " x10

u
(:in. )
u
(in. )
. c
u
3-1 699
688 1.017; 0 6.0
475
66
25 133
J-11 656
595
1.100 0 6.0
450
66 2.0 1.12
J-2 698
830; 0.840 10.6 16.6 47'5 66
94
1.21
, J-8 647 893 0725
18.2 24.2 16.00 66 16.2 0.64
J-17
662 896 0740 172 232 1340
66 18.2 0.;89
J-18 628 878 0.7
16
, 187 247 1300 66
155 073,
J-I0 913 938 0974 17 77
,361 66
35 190
J-.14
'978 1291 0756
16.1 22.1
761
66
115 1.03
J-1.3
1004 1390 0722
18.4 24.4 14.6 66
137
058.,
J-19. 932 941 0990 07 67 348
66
19
1.24
J'-20
1355 0715'
18.8 24.8
130
66
105 0.49"
J-4 1274 1511 0.844
103 16.3 4.25 66
78 171
J-9 1304 1494 0.872
85 145 360
66
50
1.46
1368 1780 0766 155 215
8.01 66 12.8 1.12
J-6 1401 1842 0760 159 219 16.4 66 10.0 0.42
1301. 1470 0.885
76 136 335
66 4.0
133
J-22 1413 1855 0761 158
21.8
119
66
73
0.4,3
*
M
.e('l"" ''/:;0
f3 = ' '.
'" J--
.**
6,
'U
K =
(w/2 + (3)cr, c
u
-124--:
TABLE 10
COMPUTED K VALUES 'FOR ULTIMATE DEFLECTION ANALYSIS -
u
BEAMSOF McCOLLISTER ij S A..l'ifDERNST I S TEST PROGRAMS
{in. -kiJ2s 2
it ..*
ComE'
M 1M
*
+ (3)
Camp. _j
Meas.
*** Beam {3
D. K
M M
e u 2
CPu
x10
No.
(in. ) (in. ) e
U
}
u-
e u In.
T-1 409 511 0.80
96 156 904
48
70
1.04
T-2 833 1059 0786 103 163 745
48 8.4- 1.44-
T-3
1221 1430 0.855 70 130 6.49 48
49
1.21
T-4 329 394
0.835
79 139 . 7 54
48
51
1.02

549
938+ 0.860
67 127
4.16 48
30
1.18
T-6
900+ '1039+ 0.866 6.4 12.4-
383
48 2.6 1.14
(4.8 + O.lw + (3)
68-2 693 655
1.06 0
54- 494 51
1.8
132
128-2 693 655
1.09 0 6.0
494-
48 1.6 1.12
188-2 693 655
1.06 0 6.6
494 45
2.0
136
248-2 693 679
"1.02 0
72 536
42 2.1
130
368"",2
693 711 0975
1,2
:. 6.03 36
'2.2 1.06
68--3

338 093 36 90
8.40 51 50 130
128--3 314 338 093 34 94
8.40 48
47
1.24-
188-3 314 338 09.3 31 97
8.40
45
35 095
248-3 314- 338 093 27
10.1 8.40 42 4.1 1.15
368-3 314 j43 092 29 113 871 36 36
1.02
.*
{3 = (1 - M 1M ) 0
e u
.*"*.
(w/2 = 4.8 + O.lw) for Ernst i s Beam.s fMc = meanmeas. mom. for group
*.**. 6.
K -
u
u - .{w/2
1
+{3 ) CPu
+ Measured moment .
-125-
TABLE 11
COMPARISON OF YIELD MOMENTS AND DEFLECTIONS FOR BEAMS
OF THE CURRENT TEST PROGRAM
Beam d Comp. Computed
M (inch-kips) M (Meas.) 6- (inches) 6. (Meas.)
Y.. Y.. Y.. ...Jl.
No. kd
xl0-
3
:Comp. : Meas. M (Camp.) Camp. Meas. 6- (Comp.)
(in. ) (in. )
cry
y
Y
J-l 10.0 4.27 0326 637 639
1.00 0.608 0.600
0.99
J ... l1 10.0
435 0330 634 648 1.02 o ~ - 6 1 5 0.610
099
J-2 10.0 4.06 0.316 645 685 1.06
0589 0590
1.00
J-8 10.0
384 0.294 607 629 1.04
0549 0530 097
J-17
10.0
389 0305
626
635
1.01
0.5
6
9 0535 094
J-18 10.0 4.16 0311 599
619 1.03 0580 0575 099
J-24 10.0
270 0.263
175 175
1.00 0.490 0.450
092
J-I0 14.0
5 45
0'.211
869 892 1.03
0394 0396
1.00
J-14 14.0
497
0.207
913 956 1.05 0386
0.400 . 1.04
J-13 14.0
471
0.196 885 940 1.06 0.366 0380 1.04
J-19 14.0
540
0.212 883 910 1.03
0395 0380 096
J-20 14.0 4',74 0.197 888 940 1.06
0368 0370
1.00
J-4 18.0 6.15 0.152 1132 1194 1.05 0.283 0.295 1.04
J-9
18.0 6.25 0.159 1182 1214 1.02 0.297 0310 1.05
J-5
18.0
'5 75
0.147 1145 1184 1.03 0.275 0300 1.09
J-.6 18.0
547
0.147 1177 1194 1.01 0.275 0.290 1.05
J-21 18.0 6.22 0.160 1198 1219 1.02 0.298 0300 1.01
J-22 18.0
553
0.148 1176 1204 1.02 0,276 0.290 1.05
Mean 1.03 Mean 1.02
. Range
1.00
Range
094
1.08 1.10
.,.126-
TABLE 12
COMPARISON OF YIELD MOMENTS AND DEFLECTIONS
FOR BEAMS OF McCOLLISTER'S 'TEST PROGRAM
Beam
d
Camp. Computed
M (inch-kips)
"l.
M (Meas.)
"'l. . .
6. (inches)
il
6 (Meas.)
"'l.
No. kd
x 10-
3
Camp. Meas. M (Camp.) Camp. Meas. 6. (Camp.)
..' (in. )
CPy
y y
s-.6 1072 349
0.248 287 280
098
0.266 0.260
098
S-7 1072 296 0.232 196 192 098 0.249 0.230
093
s-8 1072 315
0.238 198 189 096 0.255 0.240
094
T-l 1058 381
0.250 j42 ,343
1.00 0.269 0.260
097
T-2 1037 474 0322 800 805 1.00
0346 0.370 1.07
T-3
10.20
530 0365 1185 1217 1.03
0392
0.430 1.10
T-7 1058 354
0.236
335
,356 1.06 0.254 0.260 1.02
T-I0 10.58 388 0.257 348 376
1.08 0.276 0.270 098
T-ll 1058 316 0.231
349 355
1.02 0.248 0.260 1.05
T-12 10.28 4.85 0338
1006 1070 1.06
0363
0.410 1.13
T ... 13
1037 4.23 0358
1012 1073
1.06
0384 0.435 1.13
10.58
359
0.240
337 367
1.09 0.258 0.250
097
T-15 1058 361
0.268
383 406 1.06 0.288
0300 1.04
T-4 10.65
355
0.263
275 273 099
0.283 0.280
099
T-5 ,1051 4.48
0317 528
5.37
1.0.2
0341 0360 1.06
T-6 1037 505
0.410
990 985
'1.00 0.441 0.430 098
T-8 1058 0367
0.260
363 346
, 095
0.280 0.290 1.04
T-9 1072 272
0.256 200 215 .1.07 0.275 0.270 098
Mean 1.03 Mean 1.02
Range
095
Range
093
1.09 1.13
-127-
TABLE 13
COMPARISON OF COMPUTED AND :MEASURED CRUSHING MOMENTS
BEAMS OF THE CURRENT TEST PROGRA.M
d Computed Computed
M (inch-kips) M (Meas.)
Beam c c
No.
(in. )
k d(in.)
-":)
Measured
M (Comp. )
c
CPc
xl0 ~
Computed c
J-l 10.0
250 156 699 729
1.04
J'-ll 10.0
270 1.47 656 651
099
J-2 10.0
234 170 698 730
1.04
J-8 10.0 2.09 190 647 652 1.01
J-17
10.0 2.19 1.82 662 652
099
J-18 10.0 2.27
175
628 630 1.00
J-24 ID.O
155 258
202 208 1.03
J-I0 14.0
290 139 913 958 1.05
J-14 14.0 2.22
179 978
1002 1.02
J'-l,3 14.0 2.13 1.86 1004 988 098
J-19 14.0
275
1.44
932 937
1.01
J-20 14.0 2.15 1.85 969 950 098
J'-4 18.0 2.44 1.60
1274 1289 1.01
J-9
18.0
279
1.42 1304 1327 1.01
J-5
18.0 2.22
178 1368
1359 099
J'-.6 18.0 2.15 1.84 1401
1339 095
J-21 18.0
270
1.46 1301 1260 0.96
J-22 18.0 2 .. 23
179
1413 1350 095
Mean
099
Range
095
1.05
-128-
TABLE 14
COMPARISON OF COMPUTED AND MEASURED CRUSHING :MOMENTS
::1.:' ;' .
BEAMS OF MeCOLLISTERiSTEST PROGRAM
Beam d Computed Computed
M (inch-kips) M . ~ (Meas, )
e e-
No.
(in. )
.kd(in. )
xlo'-3
Computed Measured M (Camp. ) e
CPe
e
s ~ 6 ;1072
155 258 320 312 098
S-7 10.72 1.14
351
236 224
095
s-B 1072
1.46
274
220 223 1.01
T-l 1058 1.62 2.47 409 388
095
T-2 1037
2.25 178 8.33 831 1.00
T-3
10.20 '.
320 1.25 1221 1211
099
T-7 1058 152 2.63 392 388
099
T-I0 1058 1.69
237
403 415 1.03
T-11 1058 1.85 2.16 404 428 1.06
T-12 10.28
251 159
1032 1078 1.04
T-13 1037
2.20 1.82 1084
1079
1.00
T-14 1058 1.61 2.48
374 393
1.05
T-15 . 1058 170 235 431 428
099
T-.4 10.65 1.69 2,37 329 308 0;94
T-5 1051 2.48 1.61
549 545
.1.00
T-.8 1058 196
2.04 411 388 094
T-9
10.'(.2
157 255 257 259 1.01
Mean 1.00
Range
094
1.06
-129-
TABLE 15
CRUSHING DEFLECTIONS COlY1.EUTED BYIYIETHOD 1 COMPARED WITH
MEASURED VALUES
-
BEAL'\1S0F THE Cu"RRENT TEST PROGRAM
Beam
* 'x10-
3

(6 -6 )
Camp. Camp. Meas. I::::. Meas.
No.
. a K
CPe-CPy
c
c 2 e y
6 6 6 Camp.
y c e c
J'-l 66
59
0.66 1.23 119
0.64,
0.61 1.25 1.11 0.89
J'-11 66
23
0.81 1.14
83 051
0.61 1.12 1.05
094
J'-2 66
52
0.69 138
11.2 070 059
1.29 1.13 0.88
J-8 66 4.0 074
1.61 10.,0
078 055 1,.33 1.26
095
J'-17 66
.37 075
51
97 073 057 130 131 1.01
J'-18 66
30 078
1.4,4
90
0.67 058 1.25 132 1.06
3-24 66
89 054 232 149 1.23 0.49 172 175
1.02
J',-10 66
3j 077
1.18
9,,3 056 039 095
1.0J+ 1,,10
,J-14, 66 4.4 072 158
10.4
078 039
1.17 1.10
094
66
79
058
1.66
169
1.08
037 1.45 130 090
3-19
66 3.6 075
1.23
96
0.,58 0.40
098 092 094
J'-20 66
55
0,68 1.65 115 0.85
037
1.22 1.21
099
J'-4 66
73
0.61 1.45
133 078
0.28 .1.06
092 0.87
3-9
66 6.2 0.65 1.26 12.2 0.66
030 09,6 1.00 .1.04
3-5
66
107 050 1.63 167 090
0.28 1.18 1.18 1.00
,J'-6 66 10.6
050 1.69 16.6
093
0.28 1.21 1,31 1.08

,""L'
00
53
0.69 130 113 0.67 0,30
097 097
1.00
66 11.1 0.,50 1.64
171 093
0.28 1.21 1.26 1.04
......
Mean
098
K
=
> 05
e
.;.E.,;':'
( , (w \ ( ,
Range
0.87
6. -6
= Ke 'CPe 2" + a) .)
1.10
c y
-130-
TABLE 16
CRU8HINGDEFLECTIONS COMPUTED BY METHOD 1 COMPARED WITH
MEASURED VALUES -BEAMS OF McCOLLI8TER is TEST PROGRAM
Bea.m
* x 10-
3
w
*.u.
G,OIrlP'
COp1p. Meas.
A
l\1eas.
!J.. K (6 -6 )'''

ex
cP e -CPy
e
No. c 2 e y
6 6 6 6Comp.
y e c c
s-6 48
94 052 2033 154 090 0.27 1.17 1.21 1.03
8-7
48 125 050 328 185
1.46 0.25
171
1.80 1.05
s-8 48 10.1
050 250
16.1
097
0.26 1.23 1.22
099
T-1 48
79 059
2.22
139
0.88 0.27 1.15 1.20 1.04
T-2 48
19
0.83 1.46
79
0.46
035
0.81 0.85 1.05
T-3
48
13
0.85 0.88
73
0.26
039
0.65 070
1.08
T-7
48
70
0.62
239 130 093
0.25 1.18 1.19 1.01
T-10

65
0.64 2.11
125
0.81 0.28 1.09 1.09 1.00
T-l1 48
79 059 193 139 076 0.25 1.06 1.10 1.04
T-12 48
15
0.84 1.25
75
038 036 074 077
1.04
T-13
48
37 075
1.46
97
051 038 0.89 090
1.01
T-14 48 4.6
071
2.24 .'illo.6 0.81 0.26 1.07 1.00
094
48
55
0.68 2.08
115 078 0.29 1.07 1.10 1.03
T-4 48
77 059
2.11
137
0.82 0.28 1.10 1.15 1.05
T-5 48
19
0.82 1.29
0.40
034 074 072 097
T-8 48
58
0.67 178
11.8 0.68 0.28
096
1.00 1.04
T-9
: 48".:, 105C; 0.50 2.29 165 091
0.28 1.19 1.20 1.01
*
Mean 1.02
K = > 05
c
094
**
= K c (cpe -cP y) + ex) (!J.. )
Range
1.08
6 -6
c
Y
-131-
TABLE 17
CRUSHING DEFLECTIONS COMPUTED BY' METHOD 2
COMPARED WITH MEASURED VALUES -BEAlVIS OF TEE CURRENT TEST PROGRAlVI
* (cp -cp )xl0...,3 (inches)
6,
Meas.
Bea.m p
Xl
Computed Meas. e
No.
e y ,
(%)
(in. ) (in. ) (6 -b. ) b. b. b. b. Cornp.
c y y e c e
3-,1 2.0
65
1.23 66
053
0.61 1.14 1.11
O . ~ 9 7
J --11 2.0
65
1.14 66
0:49
0.61 1:._10 1.05
095
3-2 2.0
65 138
66
059
0.59 1.18 1.13 096
J-8 2.0
65
1.61 66 0.69 0.55 1.24 1.26 1.02
J-17
2.0 6.5
1
151
66 0.65 0.57 1.22 131 1.07
J-18 2.6
65,
1.44 66 0.62 0.58 1.20
132
1.10
. 3-24 0.67 8.8
232
66
135 0:49 1.84 175 095
J-I0 1.4
76
1.18 66
059 0.39 098 1.04 1.06
3-14 1.4
76 158
66
079
0.39 1.18 1.10
093
J-13 1.4
76
'1.66 66 0.83 0.37 1.20 130
1.08
J-19 1.4
76
1.2.3 66 0.62 0.40 1.02
092 090
J-20 1.4
76
1.65 66 0.83 o . ~ 3 7 1.20 1.2l
099
3-4 1.1 8.1 1.45 66
077
0.28 1.05 092
0.88
J'-9 1.1 8.1 1.26 66 0.67 030 097
1.00 1.0,3
J-5
1.1 8.1 1.63 66 0.87 0.28 1.15 1.18 1.03
3;-6 1.1 8.1 1.69 66
090
0.28 1.18
131
1.11
J'-21 1.1 8.1
130
66 0.69 030 099 097 098
3-22 1.1 8.1 1.64 66 0.88 0.28 1.16 1.26 1.09
'*
Mean I.OJ-
x ~
= [10 - -1. 75pf> 65
*
= Xi (epe -epy) (.e)
0.88
b.-6 Range
1.11 e y
-132-
TABLE 18
CRUSHING DEFLECTIONS COMPUTED BY :METHOD 2
COMPARED WITH MEASURED VALUES - BEAMS OF MeCOLLISTERiSTEST PROGRAM
* (cp -cp )xl0-
3
(inches)
6 Meas.
Beam p
Xl
Computai' Meas, e
No.
c y
C%)
(in, ) (in. ) (6-.6 ) 6. 6. 6. 6. Camp.
e y.
y e e c

093
8.4
233
48 094 o. ;27 1.21 1.21 1.00
S-7
0.62
89 328 48 1.40 0.25 1.65 1.80 1.09
s-8 0.62
89 250 48 1.07 0.26
133
1.22
092
T-1 1.4
76
2.22 0,81 0.217 1.08 1.20 1.11
. T-2
32 65
1.46 48 o . 46-. .. ' 5
0.81 0.85 1Q;5
T-j
51 65
0.88 48 0.27
039
0.66
070
1.06
T-7
1;4
76 239
48 0.87 0.25 1.12 1.19 1.06
T-I0 1.4
76
2.11 48
077
0.28 1.05 1.09 1.04
T-l1 1.4
76 193
48
071 0.25 096
1.10 1.15
T-12 4.1
65
1.25 48
039 036 075 077
1.03
T-13 32 65
1.46 48 0.46
038
0.84
090 1.07
T-14 1.4
7.6 2.84 48 0.82 0.26 1.08 1.00 0,93
T-15 1.4
76
2.08 48
076 0.29 1.05 1.10 1.05
T-4 097 83
2.11 48 0.84 0.28 1.12 1.15 1.03
T-5 19 77
1.29 48 0.48
034
0.82
072
0.88
T-8 1.4
76 178
48 0.65 0.28
093
1.00 1.08
T-9
0.62 8.9 2.29 48 098
0.28 1.26 1.20
095
*
Mean 1.03
x' = [10- .1.75p] :> 6.5
*
Range.
0.88
'.6. -6. = x' (cp -cp ) ( )
1.15
e ye y

TABLE 19
OF CRUSHING DEFLECTIONS 'BYMETHODS 1, 2., AND 3
FOR BKA..MS OF CURB.ENT TEST PROGRAM
(all deflections are given in inches)
Beam Measured Method Method 2'*;';' Method 3***
6. )/6. (comp.)
c . c'
No. 6. -6. 6. 6. -6. 6. 6. -6. 6. 6. Meth.l Meth.2 Meth.
3
c y c c y c c y c c
051
1.11 0.64 1.25 053
1.14 1.03 0.89
097
1.08
J-l1 0.44 1.05 051
1.12 0.49 1.10
097 094 095
1.08
J-2
053
1.13 070 1.29
059
1.18 1.13 0.88
096
1.00
3-8
073
1.26
078 133
0.69 1.24 1.26
095
1.02 1.00
3-17 077 131 073 130 0.65 1.22 1.20 1.01 1007 1.09
,:J-10 0.64 1.04 056 095 059 098 092
1.10 1.06 1.13
J'-14
070
1.10
078 1.17
079
1.18 1.18
094 093 093
3-13 092 130
1.08 1.45 0.83 1.20 1.23 090
1.08 1.06
3-19 054 092 058 098
0.62 1.02
095 094 090 097
J-20 0.84 1.21 0.85 1.22 0.B3 1.20 1.22
099 099 099
J'-4 0.62
092 078
1.06
077
1.05 1.06 0.87 0.88 0.87
3-9 0.69 1.00 0.66
096 0.67
097 094
1.04 1.03 1.06
J'-,5 0.88 1.18
090
1.18 0.87 1.15 1.18 1.00 1.03 1.00
3-6 0.96 1.25 093
1.21
090
1.18 1.22 1.08 1.11 1.02
J'-21 0.67
097
0.67
097
0069
099 097
1.00
098
1.00
J-22
097
1.26
09.3
1.21 0.88 1.16 .1.18 1.04 1.09 1.07
3-18 074 132 '0.67 1.25 0.62 1.20 1.16 1.06 1.10 1.14
J'-24
130 175
1.23 172 13,5
1.84 1.64 1.02
095
1.07
,*, K
[0. 004a] > 05
Mean
098
1.01 1.01
=
.c
** Xl
= [10-175pJ
:; 605
Range
0.87 0.88 0.87
H'* x = 10 in.
1.10 1.11 1.14
-134-
TABLE 20
SUMMARY OF CRUSHING-DEFLECTIONS BY METHODS 1, 2, AND 3
FOR BEAMS OF McCOLLISTERfS TEST PROGRAM
(all deflections are given in inches)
Beam Measured Method 1* Method-2** Method 3'***
6. (meas. )/6. (comp.)
c c
No. - 6. -6. 6. 6. -8 6. 8 -6. 6. 6. Meth.l Meth.2 Meth3
c y c c y c c y c c
8-6 095
1.21
090 1.17 094
1.21 1.24 1.03 1.00
098
8-7 157
1.80 1.46
171
1.40 1.65 1.69 1.05 1.09 1.07
8-8 098
1.22
097
1.23 1.07
133 1.32 099 092 092
T-l 094
1.20 0.88 1.15 0.81 1.08 1.19 1.04 1.11 1.00
T-2 0.48 0.85 0.46 0.81 0.46 0.81 0.85 1.05 1.05 1.00
T-3 0.27 070
0.26 0.65 0.27 0.66 0.60 1.08 1.06 1.17
T-7 093
1.19
093
1.18 0.87 1.12 1.26 1.01 1.06
094
I
T-I0 0.82 1.09 0.81 1.09
077
1.05 1.14 1.00 1.04
096
T-l1 0.84 1.10
076
1.06
071 096 1.04 1.04- 1.15 1.06
T-12 0.36
077 038 074 039 075 076
1.04- 1.03 1.01
T-13
0.46
090 051 0.89 0.46 0.84 0.87 1.01 1.07 1.04
T-14
075
1.00 0.81 1.07 0.82 1.08 1.19
094 093
0.84
T-15
0.81 1.10
078 1.07 076 1.05 1.13 1.03 1.05 098
T-4 0.87 1.15 0.82 1.10 0.84 1.12 1.14 1.05 1.03 1.01
T-5 0.36 072
0.40. 074
0.48 0.82
077 097
0.88
093
'1'-8 071
1.00 0.68
096 0.65
093 098 1.04 1.08 1.02
T-9 093
1.20
091 1.19 0.98 1.26 1.23 1.01
095 098
*-
= [0.9-0.04x] > 0.5
Mean 1.02 1.03 1.00
K
c
** Xl = [10-175p] 56.6
Range
094
0.88 0.84
*** x = 10 in.
1.08 1.15 1.17
-135-
TABLE 21
COMPARISON OF COMPUTED AND ULTLMATE MOMENTS -
BEAMS OF THE CURRENT TEST PROGRAM AND MeCOLLISTERlS TEST
Beam
d
f! If'
*
Camp.
Camp. -3
M (inch-kips) (lleas. )
k d (in.)
u
No.
(in. )
eu c cp x 10
Computed Measured M (Camp.) u u
u
3-1 10.0 1.107 276 475
688
739 1.07
3-11 10.0 1.107 291 450 595 617 1.04
3-2 10.0 1.234 2.22
707
829
919
1.11
3-8 10.0 1.274 1.03 16.02 894 878 0.98
3-17
10.0 1.274 1.23 1341 896 892 1.00
3-18 10.0 1.289 1.29 1302 879 876 1.00
3-10 14.0 1.094
357
361 938 943
1.01
J-14 14.0 1.222 2.02 ,762 1294 1296 1.00
3-13 14.0 1.261 1.11
1459 1391 1392 1.00
3-19 14.0 1.043 342 348 942 937
1.00
3-20 14.0 1.210 1.17 1299 1356 1238 091
J-4 18.0 1.088
301 4.25 1511 1488 0098
3-9
18.0 1.088
356 '360 1493 1356 091
3-5
18.0 1.215 191
8
v
01 1781 1800 1.01
3-6 18.0 1.255 098 16.43 1845 1740
094
3-21 18.0 1.040
352 335
1469 1363
093
3-22 18.0 1.207 1.27 11.89 1856 1650 0.89
T-l
10.58 . 1.201 1.66
904 512 512 1.00
T-2 1037 1.398 255 745 1057 1111 1.05
T-3
10.20 1.572
345
6.49 1429 1475 1003
T-4 10.65 1.201
199 754 395 391 099
T-5 1051 1.221
370
4.16
594 638 1.07
T-6 1037 1397 493
3.83 951 1039 1.09
* f' If! = 1.0 + 10 Pb
Mean 1.00
cu c
Range
0.89
1.11
-136-
TABLE

',i'
COMPARISON OF COMPUTED AND MEASURED ULTI1'vffi.TE DEFLECTIONS FOR BEAMS OF
THE CDRRENT TEST PROGRAiVJ:, AND McCOLLISTER! STEST PROGRAM
f3*)
Comp. *'., Measo Remarks
Ivleas. D.
pcr/
p
'IN U
Beam K
CPu
(.:..:. + 1
6- 6. Compo D. u '2
u u u
J-l 0
150 475
6.0 66 2.8
25
0.89
J-l1 0 1050 4.50 6.0 66 207 2.0
074-
J-2 0.56 1.11
707
16.6 66 ,8.6
9.4 1009
J-8 1.00 0 . 80 16.00 66 20.4 16.2 0.80
J-17
1.00 0080
1340 232
66 1605
,18.2 1.10
J-18 1.00 0.80 13000 24.7 66
1TO 155 091
J-I0 0
150
3.61 ,.
77
66 2.8 ' 35 1025
J'-14 0.56 loll
761
22.1 66 1105
094
J-13
1.00 0.80 14.60 24.4- 66 '18.8
137
("shear!!)
073
J-19
0
150 3.48 607
'66
23 19
U-st.irrups 0.83
J-20 '1.00 0.80 1300 24.8 66
170 105
('!shear
Ti
) 0.62
J-4 0
'150 4.25 16.3 66 6.8 6.2
091
J-9
0
150 360 145
66
52 ,
4.4 0.85
J-5 056
1.11 e8.01
215
66 12.6 '12.8 1.02
J'-6 1.00 0.80 i6.40
219
66
190
10.0 (" shear" )
053
J-21 0
150 335 136
66
45
4.0 U-stirrups 0.89
J-22 1.00 0.80 11.9
0 21.8 66
137 73
(II shearH )
0.53.
T-1 0.45 1.18
904 156 48 8.0
70
0.88
T-2 0.60 1.08 -,'745
163
48 6.3 8.4
133
T-3 00,51 1.14 6.49 130
48 4.6
49
0.94
T-4 0.65 1005
754 13
0
9
48 503 51 096
T-5 052
'1.14 4.16
127
48
29 30
1004
T-6 0.60 1008 .:,3-t83
. ,12.4
48
25
2.6 1.04
lJ
,., 'f ...
Mean
097
*
," .. ,
f3=
1(1
M
074
u
, Range
-w
133
**6.
= K CD (2" + f3) 1
u U'U"_
-137-
TABLE 23
SrJMMARY OF COMPUTED AND MEASURED MOMENTS AT
YIELD) CRUSHING) AND ULTIMATE FOR BEAMS OF ERNST r STEST PROGRAM
Beam Yield Mom. M (Meas.) Crushing Mom. Mi: (Meas .) Ultimate Mom. M (Meas.)
No. (in.-kips) Y (in.-kips) c
u
Meas. Camp. M (Comp.) Meas. Camp. M (Camp.) Meas. Camp. M (Camp.)
y c u
68-2 720 679
1.06 720 693
1.04 720 655
1.10
12S-2 715 679 1.05
735 693
1.06
736 655
1.12
18s-2 700 679 1.03 690 693
1.00 711 655 1.09
24s-2 695 679
1.02 691 693
1.00
709 679
1.04
36s-2 709 679
1.04
715 693 1.0:2 741 712 1.04
Mean 1.04 Mean 1.03 Mean 1.08
6F-2 714*
680 1.05 676 700 097
688
695 099
12F-2 720*
680 1.06 720 700 1.03 722 695
1.04
18F-2
709*
680 1.04 711 700 1.02
675 695 097
24F-2 693*
680 1.02
673 700 096 682 720
095
36F-2 656* 680
097
626 700 090 720 754 0.96
Mean 1.03 Mean 098
Mean 0.98
6s-3 274
260 1.05 320 276 1.16 366 338 1.oB
12S-3 274
260 1.05 317 276 1.15 348 338 1.03
18S-3 270 260 1.04 304 276 1.10
349 338 1.03
24S-3 273
260 1.05 311 276 1.13 351 338
1.04
36s-3 277
260 1.06 318 276 1.15 358 343
1.04
Mean 1.05 Mean 1.14 Mean 1.04
6F-3 268* 260 1.03 286 276 1.04
352 336 1 .. 05
12F-3 276*
260 1.06
293 276
1.06 346 336 1.03
18F-3 248* 260 0.96 310 276 1.12
355 336
1.06
24F-3 284* 260 1.09 305 276 1.10 340 336
1.01
36F-3 279*
260 1.07 306 276 1.11
354 340 1.04
Mean 1.04 Mean 1.09 Mean 1.04
Overall Mean. 1.04 Overall Mean 1.06 Overall Mean 1.04
*
Scaled from load-deflection curvesp. 1130 of Reference 3.
-1.38-
TABLE 24
OF COMPUTED AND MEASURED DEFLECTIONS .AT
YIELD) CRUSHING) AND FOR BEAMSO.F ERNST! STEST PROGRAM
Bea.m
No.
6S-2
128-2
18s-2
248-2
368-2
6F-2
12F-2
18F-2
24F-2
36F-2
68-3
12S-3
Load Yield Deflection
Rate
lYIeas . Camp.
0.4*
0.4 -
05*
0.25 -
03 *
0.36
0038
037
036
033
036
038
037
0036
033 .
!:s:
188-3
0.29
030
030
0.29
0027
248-3
368-3
6F-3
12F-3
18F-3
24F-3
36F-.3
w.
03 -
035*
0029
030
0.30
0.29
0.27
Crushing Deflection
Camp.. Camp. Camp.
Meas. Meth.l Meth.2 Meth.3
0.60
078
076
071
077
052
0050
0.60
0.67
0 .. 36
059
0.61
0.61
0.60
056
0,65
0067
0067
0066
0.6:1
1.96 1.04-
1.70 1.05
1.10' 1.03
1.56 1.01
1.65 0.96
1.13
095
1.2.3
1.27
1.20
1.04
1005
1.03
1.01
0096
0.63
0.63
0.61
0 .. 58
0052
0.69
0.69
0067
0.64
057
1022
1.18
1012
1.05

1.22
1018
1.12
1005
093
059
056
052
0.49
0.42
0.64
0.60
056
053
0.45
1027
1020
1.13
1.05
090
1.27
1.20
.1.13
100.5
090'
* Scaled from load-deflection curves po 1130 of Reference 3.
Ultimate Deflection
Meas. Camp.
1.8
1.6*
2.0*
2.1*
2.2*
1.6*
2.0*
2.0*
1.8
201*
4.1*
4.1
36
38
34.*
19
2.0
2.1
203
209
4.8
407
4.6
45
4.4
47
4.6
4.4
43

-139-
TABLE 25
SUMMARY OF MOMENTS AND DEFLECTIONS FOR BEAMS
SUBJECTED TO REVERSED LOADING IN THE TEST
Beam
Load Stage
Moment (in. -Kips) M(Meas. ) Deflection (inches) 6(Meas. )
No.
Measured Computed M(Comp. ) Measured Computed 6(Comp. )
J-3
Yield' (dawn)
675
.. 647 1.04 0580 0577
1.00
Yield (up) 634 647 098 0500 0577
0.87
(dawn) 711
753 095
1.11 1.21
092
Crushing (up) 686
753 091
Ultimate (down) 1000 1048 0.96 9.4 123 0'76
Ultimate (up) 1050 1048 1.00 9.8 123
0.80
J-7 Yield (dawn) 1184 . 1183 1.00 0.285 0.276 1.03
,
Yield (up) 1190 1183 1.01 0.275 0.276 1.00
Crushing (down) 1279 1417 090
. 090 1.18
076
""--.,.._r--__
... -
._-_. __ .",-.C-r:ushing.-.. ('up)_ .. _. __ . ....... 1250 1417 0.88
Ultimate (dawn) 1500 1865 0.80
70 105 0.67
Ultimate (up) 1865
J-12 Yield (down) 1234 1144 1.08
0305 0.276 1.10
Yield (up) 693 710 0.98
C:rushing (dOwn) 1269 1334 095 0,93 1.17 0.80
Crushing (up)
Ultimate (clown) 1480 1730 0.86 6.0
130
0.46
Ultimate (up)
-
,--
(
"'-
1
P
/
2
6"
12"
-140-
P
12"
, d = 10"
I I'
L.. - - - J""--Beam T -15
9' -a" Span
(a) McCollister's Test Specimen*
P/2 P/2
I
I '1
9"
1"
d 10"
11-
2
9"
3'-6" max.
(b) Ernst's Test
P/2
+
6"'
I I
- .....
)
./
P/2
l
,
-...-
12" 16",or 20
ft

,
d= 10" 14" , ,
or'18"
T
P/2
I
6"
I f
I
P/2
12" 'I
1-- 1
12 I _0" Span
I I
J./
(c) Test Specimen for the Current Test Pragram*
*'See Tables 1, 2, and :3 for description of transverse and longitudinal
reinforcement used in each of the test programs.
FIG.l
TEST SPECIMENS USED IN THE DIFFERENT TEST PROGRAMS
I
I
I
I
I
I
I
I
I
-
I
I
I
I
WEST
Roller
NOTE: All Rollers
Dia.
Moving Head of Machi.ne
J
I
-Rofler
Loading Beams -Two ... 2()"X7" I Sections
Jack: for Leveiing
Pedestal
Roller
Concrete Pedestal
Testing Machine Bed
1 ,. , 1 17 1777
..
EAST
Dynamometer
2"Diameter Ball
Test Beam
Cylindrical
Bearing
Block
FIG. 2 - TEST SET-UP FOR McCOLLISTER I SPROGRAM
I
'l-'
+='"
l-'
u
Loading Head
---- Blocking Jacks
" "- . SR 4 Load Cells, Type C,
6 WF 25 --4 d i i i' 50,000 Ib .. Cap. Each
=t
2
"
J"----' ---J
C
_
i
ft
Reinforced Concrete Heam-Column Unit
. ~ 3' -0" max. ...1
1_0"
Base of 400,000 Ib .. Baldwin
. FIG. 3 TEST SET-UP FOR ERNST'S PROGRAM
i. ~ .
..
" ~ - - I i I I I " ___ ""
..
I
l---J
..g::-
rD
a
-
. ,,:.-; ..
.,;:',",'
~
.0
H
. FIG5
I
.. \
1
I
1
Ultimate
First Crushing
--
Yield of Tension Steel
Cracking
Midspan Deflection
GENERALIZED LOAD DEFLECTION CURVE FOR AN UNDER -REINFORCED CONCRETE BEAM
I
f---I
+
+
I
o V ,.e.iIK I I I I! ........
o 0.25 0.50 0.75 "1.00 1.25
20
15
U}
Pf
:g" 10
~ 5
H
Midspan Deflection - inches
1 .. 5 2.0 2.5 3.0 3 .. 5
Midspan Deflection - inches
.0 .5
0" FIG .. 6 nPICAL CONTINUOUS LOAD-DEFLECTION RECORD TAKEN WITH AUTOGRAF X-Y PLOTTER
.. , .. .. f i ~ _ II1II ~ I11III ... .. IIIIIilII ... ... ... -
I
1-'
+=""
. \J1
I
_ .... -
I
I
f
Ul

c:..>

......t
s::
0
-.-/
-146-

1.4
c
Metb.od 1
lYlt:!"L[J,UU c:.
3

'\
\
J
;7 Neasuretl. L"c
) ./ (Crushing noted on
Gage 1 in. /from V I test. log nere - see
South Face Fig. 8 fpr phcrtograph
/' of !visual appeatrance.)
1 .0
0 .e
I/
V
/
V

;.:l. 0

.
'- Gage 1 jill. frclm
c:..>
Q)
q
Q)
p
Ul,
Q)

t)
s::
on
Q
0
.,.-!
;.:l
C)
0)
,..-i
If-i
0)
0
0
I
Face
BEAM J-2

, ,
.0

Com! uted J .

Met! ad 1 -+-----3-1 - ,... sured t.
1 I')
oc.
1.0
0'.8
0.6
0.4
0.2
0
41 Metred 2
.... Metrod 3
I /' ((".,.."C!},;-n..., on
tE 10e see
F g.8 for pI
/ A f'I vi C!11'::L1 arance )
GS! e 1 ir. =
South Face
I Gage 1 in. from
W Ncrth
BEAM J-17
0.001 0.002 0.004 0.005
Measured Strain in Concrete at Top Fiber
, FIG. 7
TYPICAL DEFLECTION VERSUS CONCRETE STRAIN CURVES
USED IN DETERMINING MEASURED CRUSHING DEFLECTION
I
I
I
I
I'
I
I
I
I
i
-147-
(a) First Crushing Beam J-2
(b) First Crushing Beam J-17
FIG. 8 PHOTOGRAPHS SHOWING VISUAL APPEARANCE OF FIRST CRUSHING
I,
I
I
I
I
I
J
I
J
I,
-148-
(a) Extent of Crushing at Ul tima.te Before Removal of Loose Concrete
(b) Extent of Crusbing After Failure upon Removal of Loose Concrete
FIG. 9 PHOTOGRAPHS SROWNING EXTENT OF CRUSHING AT ULTIMATE
FOR BEAM J-10 WITHOUT COMPRESSION STEEL
I
I
I
I
I
I
I,
I
t
J
I
J
t
t
I'
FIG. 10
-149-
(a) Beam J -17 Near Ultimate
(b) Beam J-lB After Failure
PHOTOGRAPHS SHOWING EXTENT OF CRUSHING AT ULTIMATE
FOR BEAMS WITH COMPRESSION STEEL (A = A')'
B S
I
I
I
I
I
I
I
I
I
r
I
r
I
..
(a) Beam J-2.After
(b) Beam J-14 After .F8.11ure,
FIG. 11 PHOTOGRAPHS SHOWI.M} BUCKLING OF COMPRESSION STEEL
AFrER FAILURE FOR BEAMS WITH .AI == 0.56 A
.. S s
I
I
I
I
I
I
I
i
I
I
-151-
(a) Beani J -6 After Failure
(b) Beam J-22 After Failure
FIG. 12 PHOTOGRAPHS OF BEAMS J-6 AND J-22 SHOWING SHEAR
DISPLACEMENT ASSOCIATED WITH mDE OF FAILURE
I
I
I
I
I
I
I
-152-
(a) Beams J-13 After Failure
(b) Beam J-20 After Failure
FIG. 13 PHOTOGRAPHS OF BEAMS J-13 AND J-20 SHOWING SHEAR
DISPLACEMENT ASSOCIATED WITH IDDE OF FAILURE
M
--- .-..
-------
. " , . , . . . ~
c
M
y
::E
...,.
/ I
I
d I-'

\J1
\>I
I
X
M
cr
ever C'py C'pc
Curvature - cp
FIG .. 14 MOMENT VERSUS CURVATURE RELATIONSHIP FOR UNDER-REINFORCED CONCRETE SECTION
-154- .
L-.--__ M
Section

I
. (a) CRACKING
Jcr
M
or
=r
y
. y

M
Y
(b) YIELD
M
c
(c) CRUSHING
FIG .15 THEORETICAL DISTRIBUTION OF MOMENT AND CURVATURE
AT CRACKING, YIELD, AND CRUSHING
<t
of Stub
IMy
--1...--1
:
Distribution of Moment
I'
I

j{}
B
I-'
\.J1
\.J1
S
:::::ii Measured

FIG .. 16
Measured and Assumed Distributionf3 of Curvature at Yield
COMPARISON OF MEASURED DISTRIBUTtON OF CURVATURE AT FOR BEAM J-17
WITH DISTRIBtJrION ASSUMED IN ANALYSIS FOR YIELD DEFLECTION

M
c
-156-
Face of Stub
Distribution of Moment
(Method 3) x ::: 10 in.
(Method 2) x'
...........
--..---i-- .... Computed <Pc
I .
CfJ
y
0= w/2 i (1 . - M /M )
Y c
{
Method 1 predicts peak
curvature to extend. to
this point at crushing.
at
Distribution of Curvature
-
FIG.17 COMPARISON OF MEASURED DISTRIBUTION OF CURVATURE ATCRUSHING .
FOR BEAM J-13 WrTH DISTRIBTUIONS ASSUMED IN ANALYSES
:M
c
q>c
-157-
v- Face of Stub
M
Y
x = 10 in. (Method 3)
x' Method 21
_100-.......... -..-...."""':.:-:-.. :-:-:-: ... :-:
v
L
Computed cp
c
-
~ t Method 1 predicts peak I
y ( /)/ curvature to extend to
. /. a:::: 1 - M..,. M,.. .
w/2 ~ . _ ~ this point at crushing
FIG .. 18 COMPARISON OF MEASURED DISTRIBUTION OF CURVATURE AT CRUSHING
FOR BEAMJ-14 WITH DISTRIBUTION ASSUMED IN ANALYSES
M
c M
Y
w/2
~ Face of Stub
x = 10 in. (Method 3)
' - ~ Computed cp
c
-
in. Gage Length Measure
Method 1 Predicts peak/
Curvature to Extend to
this point at Crushing.
FIG .. 19 COMPARISON OF MEASURED DISTRIBUTION OF CURVATURE AT CRUSHING
FOR BEAM J-17 WITH DISTRIBUTIONS ASSUMED IN ANALYSES
-159-
M =M
Y
c
~
I
I
+'
>:l
~
~
<Py Cj)c
Curvature - cp
FIG. 20 IDEALIZED MOMENT VERSUS CURVATURE DIAGRAM
I ~ ) M
I
1 .
~ ~
Distribution of Moment
I
I
CPc
.-
ry
Distribution of Curvature
FIG. 2l
DISTRIBUTION OF MOMENT AND CURVATURE FOLLOWING
IDEALIZED M - <p RELATIONSHIP
(a) Cracking and Position of Neutral Axis
(b) Distribution of Concrete Strain at Top Fiber
(c) Distribution of Curvature
FIG. 22 DISTRIBUTION OF CURVATURE ALONG CONSTANT MOMENT SPAN
(a) Cracking and Neutral Axis Position
(b) Distribution of Concrete S.trainat Top Fiber
(c) Distribution of Tension Steel Strain
.....
x
(d) Distributions of Curvature
FIG. 23 EFFECT OF LOSS OF BOND ON CURVATURE DISTRIBUTION
p
!
l ____ - ~
p.
w
(a) Cracking
(b) ~
L
FIG. 24 ASSUMED DISTRIBUTION OF CURVATURE FOR ANALYSIS OF CRACKING AND YIELD DEFLECTIONS
CD
cp
cr
y
I
l-'
0\
I-'
I
900
800
700
"1""1
til
PI
600

fH
.... 500


+l
:
400
fH
0
aa 300
;:j

200
X
100
0
V
o
)(
x
"
x
" J
)(
"
till ....
'300<
.
. 4JS'<

.e.
x
:f :z
- x
r :z ' 12, 000
\
x
..- + + I
X
X ...,
""
...
.... c

x x"
e

r-- x


x
.
_">I
---- .
)




x

)x
, ,
)(
P< Ji
8< X
)(
x III ' )(

.... x }
)(
X
X
X X

X )(
/
V x
/
)(
x
1
1000 2000
,.
X BE sma o Warwa Iruk'a J
CD BE 8ma 0 .. tlle C urreni Test
3000 4000 5000 6000
Cylinder Compressive Strength, f' .. psi
c
x J<


-----

)(
x x
x
f-----
Irogran:
m
-1000
8000
FIG co 25; RELATIONSHIP :BETWEEN MODULUS OF RUPl'URE AND COMPRESSIVE OF CONCRETE
I
I
u
t-.J
CJ\
f\)
I
13"
I---
r-- Face of Stub
L
6 :: 1024 cp
Y Y
CPy L-L ___________ (6s and 6F)
f--- Face of Stub
6
y
: 1075
L-_..L. ___________ __ (12S and 12F)
. I
I 9"

d
6" f--Face of stub
CD
y
,.....i ----+-_______ _______ .
6._ = 1041 :-p
y Y.
(188 and 18F)
12"
Face of Stub
= 1006 CPy
(248 and 24F)
18"
6" Face of stub
= cP
y
(368 and 36F)
54"
FIG.26 DISTRIBUTION OF CURVATURE AT YIELD FOR ANALYSIS
OF BEAMS FROM ERNST! S TEST PROGRAM
IH
f
u
Ul f
to y
f.!
+>
Cf.l
E
Y
I
,_ f == f ~ ( 112 u + 2) ( 9 ) (fu ... 1 1)].
s' y ~ + ... h f e
, . us' y
E ::: 30,000 ksi
- Esh
I :::
~ s h
Stra.in - E
FIG .. 21 COMPLETE GENERALIZED STRESS -STRAIN CURVE FOR INTERMEDIATE GRADE STEEL USED IN ANALYSIS

U
I
I-'
0\
~
B
I---
....::::::
;:;::-
1-
P
I
w 2 a;
-165-
Critical Section
.,----I-a; = f. (1 - M /M )
Y c
FIG.28 EQUIVALENT DISTRIBUTION OF CURVATURE FOR ANALSYSIS
OF CRUSHING. DEFLECTION BY METHOD 1
P/2
'1.0
0.9

0 .. 8
0.7

0.6
.....
H
0
0.5

()
lIS
r:.
t>
0.4

..c1
tf.l
0.3
0 .. 2
0 .. 1
o
o 1

... ....
...
x
....
--
I
2


""
K
0:) <: (
r
I-K = (0.9 0.04
.5
c

v

V
x

xl

,
---._--- _.
+--t---+-
e

N

...x x .,
T---

I I i
, I
---1-----
I 1
b. -L

c
r K =
1(\0 . CO. J (\ 112 + (
..... .,
. " B:Bmf
of Me ter's 1 rogrBII
CI Beam!: of tl e Cur) ent TE st Pre gram
. -
3 4
5 6
7
Spread Length into Span,
8 9 10 11 12
0: == 1. (1 -.M 1M ) - inches
y c
)(
13
FIG.?9 SHAPE FACTOR, K
c
' FOR METHOD 1 CRUSHING DEFLECTION ANALYSIS
.-
14 15
I
......
0\
0\
8
!HHU!HHH
($c - (6
C
-6
y
)
__
CD", I '
l w/2
2.
I
( a) EQUIVALENT CURVATURE DISTRIBUTION FOR 2
x
. (cp 12 wi + 8 2) = x'cp + 5 ,m
Y , . Y x c
cp 2 . 2
5 , = ..L (3'" + 12 w 1. + 8 _ x')
.x CDc 24 .
CD (q:> - ('1) )
c c y
y
(b) EQurvALENT CURVATURE DISTRIBUTION FOR METHOD 3
FIG. 30 EQUIVALENT DISTRIBUTIONS OF CURVATURE FOR
OF CRUSHING DEFLECTION BY METHODS 2 and 3
to
co
..t1
()
d
oM
-
x
...
..cJ
+'
W


w
M

...,
s::
OJ

"I'"f
&
r:iI
11

l:ttt I I I -t! I
8
7
6
5
4
3
2
1
0
0 0.5
I )(


I

.. I i1jN: I I
Lf I xl
W- x' :::1 (10
1 .. 75Ip) :;: .. 5
C I

x' =I 1. cp - !Q)v
v I.
)(
x

i
qf tl'JIe Curifent HTejst Prqgram
1.0
1 .. 5 2 .. 0 I 2 .. 5
3.0
Tension Ra.tio, p .. 10
I
3 .. 5
i
FIG .. 3r EQUIVALENT SPREAD LENGTH . FOR . COMPUTATION OF CRUSHING DEFLECTIONS BY METHOD 2
I
I
--""-
4. .. 0
B
....,
&j
8
7
6
5
CIl

Q
4
......
-
tO
X 3
2
1
0
i
I
cp 2 I
o = J... [3w +
x I cp 241
c !
!
+ 81
2
- x' ]
1 Beaml of Mfcotli+ter'S!Test
Beam of Tst

I


,. i x
)(
Ox' I 'I I; i
--t-+-! ,1- -+-__ -+-_-+ ___ +--_-1
, I
t
o 0 .. ; 1.0 1 .. 5 2.0 2 .. 5 3.0 3 .. 5 4 .. 0
1
Tension Steelj Ratio , p - %
. I
FIG.32 ADDITlqNAL EQUIVALENT SPREAD LENGTH FORI COMPUTATION OF CRUSHING DEFLECTIONS BY METHOD 3

......
0\
\0

J11
/ "'-

I
/6
(
/
10 eo 0 J--lfl"'..,1---+-__ __
< , /V "'-''-'''-
___
/'".J'B' . - "'\'
....- 0 J.,,6

"

w

Distance from Centerline--inches
FIG. 33 TENSION STEEL STRAINS MEASURED FROM
PUNCH MARKS AFTER FAILURE
Plotted K values using Measured 6 and Computed
u . --u u
;j
:x:::
H
0

CJ
cd
rx.
Q)


Ul
FIG. 34
\
K _ ,U
u - cp .e [ ( w /2)+
u
where t3 = 1 (1 - Mc/Mu) .
2 .. 4
2 .. 0
K = \105
u
+2C1'/o
lC
0.8
0.41 q
__ -L ____ ____ L ___ ____ ____ __ ____ __ __
o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
p'/p
x McCollister's Beams

Current Program Beams
<:>
Ernst I s Beams
0 "Shear" Failure Mode
PLOT OF COMPUTED K VALUES FROM ULTIMATE- DEFLECTION ANALYSIS VERSUS p'/p
u
I
&-'
-.::1
B
-172-
Moments Computed by Modified Straight Line Theory
with n = 6 + 10/f'
c
x Beams of McCollister's Test Program
Beams of the Current Test Program
. b
V
2 5 0 ~ __ -+ __ ~ ~ ~ ~ __ ~ ____ r-__ -+ ____ 4-____ ~ __ -+ ____ ~ __ ~
V
o
v
o 250 500 750 1000 1250
Computed Yield Moment - inch-kips
FIG. 35 PLOI' OF MEASURED VERSUS COMPUI'ED YIELD MOMENTS
... 173-
xBeams of McCollister's Test Program
-Beams of the Current Test Program
,
V
!
{f
I
0.6
I
i
V
/
,
I
I
!
f
1
/ V
;/
0.5
J
I
i
I
+ l ~
~
V
1/ /
I
I
I
Uj
, .
;:
V ~
'/-
.8
!
r-.-- lot
(,)
1
~ 0.4
I
~
V
,
I
IA
d
I
0
I
'r! I
Y/
/
+'
I
(,)
I
~
v
I
// ~
0.3
I
.-f-.
v
I
!(l;
~ /
~
I rO
~
rl
(!)
-M
(1
y >-t
rO
/
(!)
0.2 ~
~
~
V
ro
oj
~
/
~
/
0.1 /'
'/
/
0
V
0 0.1 0.2 0.3 0.4 0.5 0.6
Computed Yield Deflection .;.. inches
FIG .. 36 PLOT OF MEASURED VERSUS COMPUTED YIELD DEFLECTIONS
-174-
x Beams of McCollister's Test Program
Beams of the Current Test Program
1500
I L
V
I
L ~
~
+ 5%
~
V/ V
v/
1250
~
v"-
-- 5%
/
~
V
. .
L
~
V
/
~
V
A
F/
m
Pt
-r-!
~
I
~
1000
CJ
d
Ti
+>
d
Q)
~
::s
750
~
-rI
~
til
~
V
/
L ~ V
:s
H
0
rd
Q)
500
~
co
a:S
cu
~
/ ~ v
V
250
o ~
o 250 500 750 1000 1250 1500
Computed Crushing Moment - inch -kips
FIG., 37 PLOT OF MEASURED VERSUS COMPUTED CRUSHING MOMENT S
2 .. 0
to
1.8
()
s:::I
or-!
d
1.6
0
"f"'i
.+l
C)
1.4
tV
q


1.2
-1j
-

til

1.0
u
rd
Q)
S
0.8
en
cx3

0.6
0.4
0.2
0
0
-175-
Crushing Deflections Computed by Method 1
with K = (0.9 - 0.04 a) 5: 0.5
c
where a = 1. (1 - M 1M )
y c
Beams of McCollister's Test Program
" Beams of the Current Test Program
/
+10%
V
/y
/
/ V
V
/
V
V

V'
..... -
8/


y:


/

/
/

/
/

/
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1 .. 6 1.8 2.0
Computed Crushing Deflection - inches
. F'IG. 38 PLar OF MEASURED VERSUS COMPUTED CRUSHING
DEFLECTIONS BY METHOD 1
til
.8
0
d
.,.;
d
0
.,.;
+>
(J
()
q


.,.;
.s:=
en
e
0
rd
Q)

til
a3

-116-
Crushing Deflections Computed by Method 2
with X, = (10 - 1.75 p) > 6.5
x Beams of McCollister's Test Program
Beams of the Current Test Program
2 .. 0
1.8
1.6

1.2
1.0
0 .. 8
0.6
0.4
0.2
0
7
/
[7
/
V" 7
l7
V

7
v/
/
/'

V
--- lCYf,

A


l/ -/

j?
/
If;
/'
/

V
/
,v
V
. 0 0.2 0.4 0.6 0.8 1.0. 1.2 1.4 1 .. 6 1.8 2.0
Computed Crushing Deflection - inches
FIG.39 PLOT OF MEASURED VERSUS COMPUTED CRUSHING
DEFLECTIONS BY METHOD 2
2.0
g)
1.8
Q)
~
tJ
;1
1.6
s::
0
"r"I
1.4
+>
tJ
~
-r-f
G-t
~
1.2
a>
.,-t
..Q
1.0
co
~
H
0
'd
0.8
>
g
CIl
as
~
0.6
0.4
0.2
0
-177-
Crushing Deflections Computed by Method 3
with x = 10 in.
x Beams of McCollister I s Test Program
Beams of the Current Test Program
/
V
L
/
k<
/ ..
V
:/
L
+ 15%
~ V L
V
I
~ V /
V
V" fJ
V
"-~ - 15%
/
l0.
/ v:
V'
/x: XV
V
L 7/
~
V
/
, ~
V
Vv
o 0.2 0.4 0.6 0.8 1.0- 1.2 1.4 1.6 1.8 2.0
Computed Crushing Deflection - inches
FIG.40 PLOT OF MEASURED VERSUS COMPUTED CRUSHING
DEFLECTIONS BY METHOD 3
... 178 ...
)( Beams of McCollister's Test Program
Beams of the Current Test Program
2000
/
co
1800
~
~
I
..c=
C)
1600
s::I
"1""'1
~
1400
~
:::i:
t)
1200
j
...,
S
1000
rd
(I)
H
-800 ::s
UJ
(11
t>
::E:
600
400
200
0
/ V
~ 10%-
IV
V?
b/
/ V
V
V/
V
'-- ... ]
10<,11
/ v-/
A ~
/
r . ~
V
/
~
/
A
~
V
/
VV
V
r
o 200 400 600 800 1000. 1200 1400 1600 1800 2000
Computed Ultimate Moment - inch-kips
FIG .41 PLOT OF MEASURED VERSUS COMPUTED ULTIMATE MOMENTS
22
20
18
til
~
16
0
d
-rl
d
14
0
.,..;
~
c:.>
OJ
r-I
12
ct-t
~
Q)
~
10
oj
~
~
r-f
:;:J
8
rO
Q)
H
;:s
6
ta
cd
(I)
~
4
2
o
/
/ ~
V
M
o 2 4
... 179-
Ultimate Deflections Computed from:
/\ = K Il) (w/2 + ~ ) 1
u uu
x Beams of McCollister's Test Program
., Beams of the Current Test Program
o Beams of the Current Test Program Involving
Shear in the Mode of Failure
v
V/
/
V
/
/

/
V
V 1/
~ 20%-
V
/

V
V L /
V V
V
0
/
/
/

/
V
//
0
~
[;/
/
.....,
K/
/
r - - 2 ~
V /
~ /
0
/
.
V. /
b{
V
V
6 8 10 12 14 16 18 20 22
Computed Ultimate Deflection - inches
FIG.42 PLOT OF MEASURED VERSUS COMPUTED ULTIMATE DEFLECTIONS
1130
20
10
III
II..
~
Z 0
-
0
"" 0
..J
..I 50
""
I-
0
I-
o
-180-
JOURNAL OF THE AMERICAN CONCRETE INSTITUTE
I INCH
GROUP I 5
GROUP IF
-,-r--I
I INCH
GROUP 2 5
GROUP 2 F
{VCL
245
Y
eL
185
C)CL
125
2 4 0 ~ 4
OULECTION I N INCHES
Fig. 6-load-deflection curves
June 1957
Group 1
Beams
[Over-rein ....
forced and not
included in
this study]
Group 2
Beams
Group 3
Beams
FIG. 43 .LOAD-DEFLECTION CURVES FOR BEAMS OF ERNST'S PROGRAM
50
45
40
35
to 30
Pi
"rl

25

0
H
20
15
10
5
o
te

"
rf


"'-

..cI
r-J -4
Ul
::s


,
"'"

"'-I-J -9
\

I


.,. ...
'--J -J 0
"\
-,
K
f-J-l
AV
............
(--
rr-

Beam
d f'

-- c

r-J -11
J-l 10.0
4930
I
"'I'
J-11 10 .. 0 4110
J-10 14,,0 .
3590

J4 18 .. 0 4820
J-9 18 .. 0
4190
,
I
o 1 2
3 4
5 6
7 8
9 10 11
Midspan Deflection in Inches
FIG. 44 .EFFECT OF DEPl'H FOR BEAMS WITH P I jp = 0
f'
-=Ella l
47,600
46,900
45,100
44,900
47,000
12
u

. OJ
J-A
I
55
50
45
40
35
to
p.
'rl

30

0
25 H
20
15
10
I

5
0
0
J-5 :-


I
----
V .......
bO

V roo
i-
. - _. _ ...
::s
/'
lj
u:
/ 1!IiII
r&.
---1----r---
..lit.
!

-
\


.

,.....


I
.
I .

r
--+---t--
-,-
J-2
L .J..
-----

\
. -----
.....
-t-+
-i
/1 I 1
i
i
Beam d f' f

c y
I
I
3-2 10.0 4080 48,000
I
I
!
J-14 14.0 4500 47,100
I
J-5 18.0 5000 45,100
I
____.l_
I I I
i
I I
1 2
3
4
5
6
7
8
9
10 11 12
Midspan Deflection - inches
]'IG .45 EFFECT OF DEPTH FOR BEAMS WITH P I /p :: 0 .. 56

.....
13 14
I
S;
f\)
B
M%i
-
WT'M' '"
55

te

..........

50



t-3 -6
'" gf
/

oM

45
.c1
:3
V
"'Q

H

/
L...,..............

""""" 40
/



"

t- J -13
35


P""'"
rJl


Pi
;l 30


/
r- J -.1"/
I
,
lY
25
1-'
A
'"
H

W " -8 and 3 11

I
--
.....
CD
20 \.)J
r

\
I
Beam d f' :f
15
c y
....
3-8 10.0 4680 45,400
10
J-11 10.0 , 3900 46,900
3-13 14.0 4800 45,600 I
5
J-6 18.0 5160 46,200
0
0 1 2
3
4
5
6 7
8
9
10 11 12 13 14 15 16 11 18
Midspan Deflection - inches
FIG .. 46 EFFECT OF DEPrH FOR BEAMS WrrH p I jp = 1.0

50

45
40
35
J -22-
[>-


!



V
.,
0

--'"'
fe, ... ..,....

...,....,-
I-J -2]
"\
-
LY

\

---
"
"


tl]

rl
30 .!4
rg
25
0
H



K
-J-20



f-J -15
"
8

. +:-

20
Beam
p'/p d
f' :f
c
l.
.J-l9* 0 14.0
3900 45,800
15
J-21* 0 18.0
4350 47,600
J-20** 1.00 14.0
4380 45,800
10
J-22** 1.00
18.0 4420
46,200
#. 6'
* 2 - U-Stirrups at in ..' Spacing
5
**#2 Closed Stirrups at 6 in. Spacing
o
o 1 2
4 3 6 5 8 7
9
. 10
11 12
Midspan Deflection -.inches
FIG. 47 EFFECT OF DEPrH AND COMPRESSION STEEL FOR BEAMS WrrH 12 STIRRUPS
..
30
Ul


J"-2 -

J..
.L
t-J -17
"
.,

,
25

H

r--
1\
"-
J
til
P

J-8 ax d J-l"
J-B - P4
i\
rt


20
. j
,
F-

\
r
r-J-B
rd
p'/p


I
J-ll Be8JJ:l
fl
f CP
0
c y \Jl
H 15 I
"
J-l 0 4930 47,600
f-+
J-ll 0 4110 46,900
10 - ... -
J-2 0.56 40Bo 48,000
J-B 1.00 4680 45,400 -
i
5
J 1 .. 00 3900 46,900
0
I I I I I I I
0 1 2
3 4 5
6
7 8
9
10 11 12 13 14 15 16 17 18
Midspan Deflection - inches
.
FIG.48 EFFECT OF COMPRESSION STEEL FOR BEAMS WITH d = 10 in.
45
40
~
35
:a
;
H
0
30
t:Il
. ~
~
25
~
V
Ef""
~
0
H
20
15
10
5
o
o 1
t 1t1ma1 e
I
"
. J -;L3
t:>---
I"""'"'
""
~
~
-j
, , ~
~
~
~
~
\ ~ ....",,- ~
~
~
~
"-
f- J -14
" ~
~
..,
~
f-J-IO
'\
Beam p'/p f' f
c
1-
J-10. '0
3590 45,100
J-14 0.56 4500 41,100
J-13 1.00 4800 45,600
I I \ J I I I
2
::5
4
5
6
7
8
9
10 11 . 12
13 14 .15
Midspan Deflection '- inches
FIG.49 EFFECT OF COMPRESSION STEEL FOR BEAMS WITH d = 14 tn ..
16
8
~
0"\
D
55
50
45
40
35
~
Ul
Pi
.,....
~
30
~
25 0
H
20
15
10
5
o
o
Ultimate
-
~
--"
"-~ J - 5
~
!P"" .
"
~
~ ~
~
~
"
~ J - 6
. ~
~
bO
d
/
~
..-t
W
~ ~
::s
~ ~
~
"-
I- J-4
"-
"
~
H
#
,
~
,
"'
~ .. "'-
r- J-9
,
"
1\
Beam pi/I! f' f
c
Y:.
J-4 0 4820 44,900
J-9 0 4190 . 47,000
J-5 0 .. 56 5000 45,100
J-6 .
1 .. 00 5160 46,200
-
1 2
3 4
5 6
'7
8
9
10 11 12 I 13 14 15
Midspan Deflection - inches
FIG50 EFFECT OF COMPRESSION STEEL FOR BEAMS WITH d = 18 in.
16
I
b;
-:]
8
til

45
40
35
30
:Q 25
rg
.s
20
1"
'"
10
5
, 0
o
J-13
D-
-Y
"

-


"",Z,

.,.--

.......... ,,---
\ ''iii .-
-

... ..",.. ....
J-20

i'

"",,--""
p
..... -
_.fIIIIIII>""" -

::. .....
v




1\

L
J-19
Beam P'/P
f' f
c
l.
J-1O* 0
3590
45,100
J .... l9** 0 3900 45,800
J-13* 1.0 4800 45,600
J-20** 1.0 4,380 .
45,800
* #3 Stirrups at 6 in.. Spacing
** #2 U-Stirrups at 6 in. Spacing
1 2
3
4
5
6
7
8 9' 10 11 12 13 14
Midspan Deflection ... Inches
FIG. 51 EFFECT OF VARIATION IN TRANSVERSE REINFORCEMENT FOR BEAMS WITH d = 14 in.
15




rJJ
A
.,...
.!4
11
.3
55 rl
_l
5
0
1,


\
35
30 1-1 I
25
Beam
p'lp f'
f
I-
C
Y
J-4*
4820
44,900
0
J-9*

4190 41,000
J-21.**

4350 41,600
J-6* 1.0
5160 46,200
1.0 4420
46,200
20
15
*
#3 Closed Stirrups at 6 in. Spacing
** #2 Closed Stirrups at 6 in. Spacing
10
5
o
o 1 2 3' 4 5 6. 1 8 9 10 11" 12 13 14 15
Midspan Deflection - inches
FIG 52 E:lrFECT OF V ARIATION IN TRANSVERSE REINFORCEMENT FOR BEAMS WITH d = 18 in ..
I
5;
\0
a
(8.) Crusbing Af'ter Failure for B e ~ j -19
(-p) Crushing After Failure for Beam J ... 6
; . ~ . .
FIG. 53 PHqTOGRAPHS SHOW:tNG C'RUSHING PATI'ERN AFTER FAILURE FOR
BEAMS OF THE CURRENT PROGRAM HAVING U-STIRRUPS
.... --
25
20
~ 15
-.-I
.!4
'rd
r.U
o
H
10
5
o
I
~
o
- ~ -------

.J
~
L
.
V
/'
/ /
~ /
~
/
/
/
~
V
./ AP
~
//
/. ,
~
~
V
/
V
L
17, V
//
V
/'
V
. 0.10 0.20 0.30 0.40 0 .. 50 0.60 0'.70
Midspan Deflection - inches
FIG. 54 LOAD DEFLECTION Cill1VE FOR BEAM 3-17 SHOWING CHANGE IN SLOPE OF
REPEATED LOADING ClJRVE PRIOR TO INITIAL YIELD OF THE TENSION STEEL
I
I-'
)3

-
'1.0
0.9
0.8
0
0,,7
CD
-......
oM
CD
0.6
tQ
rJ.l

0.5
8
tJ)
~
0 .. 4
H
S
0.3
~
0.2
0 .. 1
o
~
~ I.a
C!

~ .
. )\
II fl

.,
~ ~
II

,

6.
1
\I
. ~
-192-
6-
P
Deflection


......
Ie
0
A 'II
....
1'-

.. ".
..
.,
o 0.1 0.2' 0.30.4 0.5 0.6 0 ~ 7
DAMAGE RATIO - 6i/t:p
j

0
.,


~
4D
49
.,
-.ft
~

4
..
~
0.8 ,0.9, 1.0
FIG.55 EFFECT OF DAMAGE ON STIFFNESS
I
I
I
-05
Cycle (a)
/
/
------ +/_---
"
-193-
Cycle (b)
-0 5 -1 0
+0 5
Cycle (c)
- . ches
FIG. 56 LOAD VERSUS DEFLECTION FOR BEAM J-3
--r-----

."

-.--


---rJ? I. on ....
-4
I

, VI j
f7 I
,'I . -
, .... ",
V1
:- I
/t-
CYClE
r,y
V'"
I .
(f)
I
I I
'... ,"() A
I

v-
,'-
_Cycle

V

(d)


(e
" I
,
I J I
d De!] ectitlf
V
hee .
Va.
l
I
!
UPW8.1
- inc
I ",.
I 5 -4 0 -3. 5 -3. o l' -2. o -1.
5 -lA
/5
V
/
+0.
5
o +1. 5 . +2.

I
I
nward Def
I
/
I
.
c::
71
I ."
v/
V
,
,
j
I
"'11"'\ / .
,.
I'
"(d)
....
'/
/
/ ;t
I
1-3\_
./
f
I
1\ ..... ' 7


r.-{e)
J
/
r-{e)
I
L/ I")ft I
V

_v
--
r---1 I



r'\C D


"-./


21'\
.J\J
FIG. 56 LOAD VERSUS DEFLECTION FOR ,BEAM J -3 (CONTINUED)
...sI!IIiJ

If/

CYClE

J
/
5 +3.
tion .
---
!If""- -

I "
j
1
i'-
r.-(e)
rt
o +3.5 +4.
inches

o +4.

I-'
5 '.
D
I
/ I i
.-+ II I I I t - - - - ~ - - - - - I - i f - I.' I ~
I I
t----L----
! I
,I + , II I t----I
I
I I
-i--i
BEAM IJ-3 (qoNTIN$n)
I-'
'D
Vl
8
-0. 5
Cycle (a)
-0.25
5
-196-
... 0.25
+0.25
Cycle (b)
ycle
(e)
FIG. 51 LOAD VERSUS DEFLEGTIONFOR BEAM J-7
"{'"\
,)V

r


,
fJl
/.-L IJ Pc j.A

'TV
,11

/.;
"r
I
I
II
} rq
7)0/
,
I
.
II
l

I
j
I
90 I
,
I,
I
,
I
,


i'
I
,
I
L rl 1 n J
,
.
in./

!I
,
F
I
] ward I Deflec . n -
B
,
1
D -;

,j -
,
u
(
, ,
"t-.L +c:
[5
+4 +5 +b +'( +b +;1 +10
6
,
DownWI p.rd De lectic I>n - il ches
8
,
,
11
I
I
I
V
6
,
I
I
,

I
I
:g
/
Cy( Ie (f
,
----- -
,
I
,
8
:,n ..rr:!
Cy( Ie (g
...
I
I
Ll
I(
Ie (h
J
-
--.. ... --- Cy(
, I
,
J."

/'


f;
c:;n
FIG .. 51 LOAD VERSUS DEFLECTION FOR BEAM J-7 (CONTINUED)
i" ... __
L
+1.,
1---
1
B
B
I
I
D
I
I
,

I
I-'
\()
-:j
I
to
ttl
"
M

m

0
..,-t
to
til
"
F-1

u
;
0.5, 1 0
1.5 2.0 2.
1
" ,.0 3-51 4.0
Strait - % /

2 0
30
4 0
5
0
I I J /
I :
, .
, / /

II ....
ABOVE DATA TAKEN FROM REFERENCE ( 8 )
RECORDED STRESS -STRAIN DIAGRAMS
FOR SoA.E. 1030 HABBON STEEL
ANNEALED AT 1675 F . - FIRST 1/2
CYCLE IN TENSION FOLLOWED BY
CYCLIC STRESS: 50,000 psi.
FIG. 58 BAUSCHINGER EFFECT IN S.A.E. 1030 STEEL
WITH FIRST HALF CYCLE IN TENSION
I
4.
I
-199-
1::.1\
vv
Q-

----::

f
1
,.,-
V--


.Iv
,
"I""i


to I
..!lQ I
/ . \,....

V
/;/
'TV
f
OJ
OJ
/
v
'71\
I //
Cf) .-IV
J

I
0
,
I "I""i r,r,
,
.
, .... \oJ
I
I
I
v
I
I
8
J l 11\
I
I
i
I
I
.... v I
I
i
I
I'"\J
b.
2"\l 3 ;
P p. D 5.P 2 5
I 2 0
1105 1 .. 0 0105 tv 1:0
1
1.0
Stra
- %
I
,
I ,rl I I
,
I
.
I
I
,
I
I
7"
/
I
I I I .
r.;J.J-z.c
I
I
,

1/
I
,
,
),... IA
L'
V
-
V /"
Vco ""
v
,/'
Cf.I

/ ....
c .


I


;
I
I
..
0
l...----
Y4-

u



ABOVE DATA TAKEN FROM REFERENCE (8 )
RECORDED STRESS -STRAIN DIAGRAMS
FOR S.A.E. 1030 STEEL
ArffiEALED kT 1675 F - FIRST 1/2
CYCLE IN COMPRESS ION FOLLOWED BY
CYCLIC STRESS: 56,000 psi.
FIG. 59 BAUSCHINGER EFFECT IN S 8 A E. 1030 STEEL
WrrH FIRST HALF CYCLE IN COMPRESSION
I
I
-0. 5
U
-2
pward
.0 -1.
J
(
dl
l
II
--
Cycle (a)
I I


Deflec

-0.5
V
J
I
If
Ja
..-tII8fIfII'1""'"
... 200 ..
-0. 5 -0.25
Cycle (b)
.

Cycle (c)
....
til -';V
Pt
-r-!
40
--
_1..---."


ir
-,
,,"
yr
J 7n .,,/
rO ../V
:, ,

f
Cycle

Cycle
H
,... (d)
- (e) a:J
:LV

8 r)t"'\
8/-VfIV-
Cycle
1
'1/
(c)
p
,;P , r.
l!
.... '"
I.:
I 1
in.
I I'
,
'" I 0
50 +1. o +1. e +2. o +2.5 +3.0 +3.
" +
' I
Dc
wnwaj
Defle ction - incl: es
, II
I ., ""' til
- "
I j .J..V
V
" I

, I
1/
I
'1
"''''
A
{'

:;:: ..-./'
---
:--'" H
'If)rO
/ ;


\t('\
FIG.60 LOAD VERSUS DEFLECTION FOR BEAM J-12
5
c::::n
./'"
-
Y
--

til

",.....
Pi Ion
"TV
... ---
r
r
"



I
,"

6
!
I
rei 7("\ I
../V
I
/
I
A /
H I ,

rd
1/
I
H r'\("\ If


.
r
if
,

I
J
I
I
J
I
J


.... '" I
.
I
v
/
Upwar< DeflE - in.
I
I
I
J
-1
5/
o -0
5 I'
I
I

,
81(
I
+0 5 +1 o +1

+2 0 +2 5 +3 0
+3 5 +4 0 +4
5 +5 0 +5 5 +6 0 +6 5 +7 0
+7 1"5
I
I
IA I Do' Defle( tion r- inch i'lS ....
/I.
., ...... I

I II .I..LV til
,

I /'
Pi I
rl
I

/
/.,
, B
,.,'I
I
I "'..... B
,
1/
Ic..V
,,' I rd
- Cye e (d) I
--g-
-- ----
I
----
I
,
.. H
Cyc e (e)
---
-.r>.
--
..JV'g
ro

I", g.
"'t"v
FIG .. 60 LOAD VERSUS DEFLECTION FOR BEAM J-12 (CONTINUED)
... 202 ...
(a) Beam J -i2 Af'ter Failure (Load Reversed)
(b) Beam J ... 5 After Failure (Load Downward)
FIG. 61 PHOTOGRAPHS SHOWING DIFFERENT lDDES OF FAILURE FOR
COMPANION BE.A1 WITH AND WITHOUT REVERSED LOADING
45
40
35
30
CIl
25

"d 20
c:d
o
H
15
10
5
o
l
o
Beam Loading p'/p
fi
f
c y
.-
J-B Down 1.00 4680 45,400
J-3 Down andL Up 1.00 4900
.
.......



,

J.;.3
j
I
- ... -

I
. J-B-U

------
------
..dIIfIJ


1
,
J
/
/
JV
.-
2
3
4
5
6 7 . l' 8 9 10 11 12 13
Deflection - inches
14 15
FIG. 62 EFFECT OF REVERSED LOADING ON DUCTILITY .AS SHOWN BY COMPARISON OF THE
DOWNWARD LOAD-DEFLECTION CURVES FOR BEAMS J -3 AND J -8 WITH d = 10 inG
,
i
16 17
I

\J.l
I
55
50
45
40
.. /
35
w
~ 3 0 ,
~
~ 25
.s
20
I
,15
10
5
o
o
,
~
~
,-
'"
~
~
~
1\
L ~ ... 6 "'-
~
V
~
~
/
V J-7-
~
~ ~
.'
~
1
~
~
"
....
Beam Loading. p'/p
f' .
f
c
l
J-6 Down 1.00 5160 46,200
J-7 Down and Up 1 .. 00 4450 46,500
~
..
\
2 . 3 4 5 6
7
8
9
10 11 12 13 14
./
Midspan Deflection - inches
FIG.63 EFFECT OF REVERSED LOADING ON DUCTILITY AS SHOW1'l BY COMPARISON OF THE
OOWNWARD LOAD -DEFLECTION CURVES FOR .BEAMSJ -6 AND:,J -7 WITH d :: 18 in ~
"
l
la5 .16
I
~
.r::-
D
55
50
45
40
35
co 30
~
~
25
11
o
...:t
20
15
10
o
-
~
~
,.......
IJ
~ 3-5
. ~
~
V
-
~ ~
~
~
~
"
/
3-12
i\
~
/
\
Beam Loading pl/p f' f
c
l
-
3-5 Down 0 .. 56 5000 45,100
J-12 Down and Up 0.56 4550 45,100
1 2
3 4
5
6
7
8
9
10 11 12 13 14 15
Midspan Deflection - inches
FIG. 64 EFFECT OF REVERSED LOADING ON DUCTILITY AS SHOWN BY COMPARISON OF ' l ~ H E
DOWNWARD LOAD-DEFLECTION CURVES FOR BEAMS 3-5 AND 3-12 WITH d = l8in
B
~ .
'VI
I
APPENDIX A- -MATERIALS , FABRICATION, AND TESTING DETAILS
29. Materials
(l) Cement - Type III Portland Cement was used for all the test
beams of the current test program. The cement was purchased in paper bags
and stored under proper conditions. MarCluette brand cement was used "for
Beams J-l through J-17 and Atlas brand cement for the remainder of the beams.
(2) Fine and Coarse Aggregate -The fine aggregate was Wabash River
torpedo sand with a fineness modulus of 3.1. The coarse aggregate was Wabash
River gravel ofl in. maximum size. The specific gravities "We're 2.65 and 2.70
for sand and gravel respectively. The a-bsorption of both fine and coarse
aggregate was about one percent by weight of the surface dryaggregate.
The origin of these aggregatesisa' glacia.l outwash, mainly of the
Wisconsin glaciation. The major constituents of the gravel were limestone and
dolomite,with minor Cluantities of Cluartz" granite" gneiss" etc. The sand
consisted mainly of Cluartz with the coarser fractions similar to the gravel.
(3) Reinforcing Steel -The' intermediate grade::deformed ba.rs used
for -reinforcement met the reCluirementsof.ASTM Designations A15-39 and
A305--50T. Except for steel used in Beams J-l through J-3, all the bars in
each size were from one heat in order that the properties might be maintained as
uniform as possible . The bars were purchased in 28-ft lengths j two 13-ft
lengths for use as reinforcement and a 2-ft test sample were cut from each bar .
Properties of the bars used in each beam as determined from tension tests on
these samples are given in Table Al.
The tension tests were made in a 120)000-lb capacity Baldwin
Southwark Tate-Emory Hydraulic testing machine. Strains were measured with
-206-
-207-
an 8-in. extensometer and recorded with an automatic recording device up to
about 2 percent, and dividers were used to measure strains within the 8-in.
gage length beyond this point in order to obtain the complete load-strain
curve. The completestress .... strain curve (based on nominal area) shown in
Fig.Al is typical for the intermediate grade reinforcing bars used in the
current program, and the generalized stress-strain curve of Fig. 27 which was
used in the analysis follows the form of this curve very closely.
Compression tests were made on pieces of No.6 and No.8 bars taken
from the same heats as those used in the beams of the current program. The
ends were threaded and screwed into steel plates to achieve fixed ends, as
shown in thephotographsofFig.A2. The lengths between fixed ends. were :4,
6, and :S- 'in. , and
a
. 2-ln,' gage length multiplying extensometer mounted as
shown in Fig. A2a was used to measure strain.
The primary object of these compression tests was to check the
assurnptionthat the stress-strain curve for a reinforcing bar is the same in
compression as in tension, . particularly the strain hardeningregion of the
curve. Figure A3 shows a typical stress=strain curve taken in these compres-
sion tests compared with that ~ e a s u r e d in a tension test using pieces cut from
the same bar. These curves indicate that strain hardening began at a slightly
smaller strain in compression than in tension,but otherwise the early part
of the curves were ,identical as assumed. Because of the uncertainty of the
strain in the compression steel at ultimate the effect of strain hardening was
neglected in the analysis, but in several cases the strains measured at ultimate
were in the order of the strain at beginning of strain hardening. With more
precise expressions for the ultimate strain at crushing of the confined concrete
core at failure,the use of the complet;= stress-strain curve including the
effect of strain hardening would be justified.
-208-
The stress-strain curves of Figs . 58 and ,59 taken from Referenee 8
show the Bauschinger 'effect for 'annealed S.A.E. 1030 steel, and it seems
reasonable to assume that the same behavior would be observed for 'intermediate
grade reinforeing bars. These curves also indicate that the initial half cycle
in tension gave the same stress--strain curve as that in compression, with the
sarhestrain at initiation of strain hardening.
30 .. Fabrication and Curing Procedure
Preparation of Steel Reinforcement
After the steel had'been cut to 13-ft . lengths for use in the speci..m.ens,
.the barswe.re prepared for mounting of electrical' strain gages at desired points ..
For Beams J-l through J-IO, a slot 1/8-in.'wide and 3/32 in. deep was milled in
the bar 'fora lengthof3in.j and a Type A 7-4 SR-4 electric -resistance strain
gage (1/2 in. gage length), torimmed to a .width of'i/8 in., was rnounted in the
bottom of the slot . After the leads were soldered to the gage , one lead wire
was carried out each end of the slot and the slot was filled with epoxy cement
to vraterproof the gage as well as secure the lead wires against being 'pulled
out during 'casting 'of the beam.
The reinforc-ing barsforBeamsJ-ll through J-24 'were .handled in a
slightly different way since the Type PA- 3 SR-.4post-yield gages (3/ 4 in .. gage
length) used on these bars could not be trimmed to the l/B-in. width. The
deforrnation lugs were ground flush with the bar in the region v.rhere a gage was
to be mounted, and a medium grade of emery cloth was used to smooth the surface
of the bar and remove'mill scale. The -gage was then mounted onthe curved
surface of the bar. An asbestos . roofing 'compound which remained elastic was
used to protect the gage and at the same time leave the lead wires and gage
free to move -slightlYjbuta covering of epoxy cement was placed Qverthissoft
-209
material to protect the gage from moisture damage and also to keep the leads .
from being pulled out during casting. Early tests on tens ion specimens
showed that solid epoxy would waterproof the gage, but the material was so
stiff that the post-yield gage was pulled off the bar long before the 10 per-
cent strain capacity was reached. Post-yield gages gave strains well out into
the plastic range of behavior of the beam,and, in many cases) almost to
ultimate when the described above was used.
Checks were made on waterproofing of the gages by immersing the
prepared bar in atTough;: of water for a period of Heveral hours ,following
which the gage resistance and resistance to ground were checked. The procedures
described above gave excellent Tesults,both on tension test specimens which
were tested in a testing machine for preliminary checks and in the' beam tests
themselves.
Punch marks were made at 2-in. centers along the rib of one of the
tension bars in each beam. A measure of the magnitude and distribution of
strain along the steel in the vicinity of the stub at ultimate could be
obtained by taking final readings from these punch marks at the end of the
test. Figure 33 shows typical distributions of strain after failure measured
in this manner .
. The stirrups were made of No. 3 deformed intermediate grade bars
welded into a closed rectangle as shown inFig.A4. These stirrups were
spaced along the prepared bars at 6-in. centers and securely wired to the
longitudinal steel to form a cage of reinforcement. This cage was then placed
in the .fOrm and chairs beneath the' bottom bars assured the desired position
of steel.
Casting of Concrete
A non-tilting drum-type mixer with 6- cu. ft capacity was used for
mixing all concrete. Two to five batches were 'each beam,
-210-
depending on the size} and the concrete was placed with the aid of a high
frequenoy laboratory-type internal vibrator. ,A butter mix was used to condi-
tion the mixer prior to the first batch, but the strength of 'different batches
having the same nominal proportions varied to some extent.
The consistency was difficult to control because of the widely
varying moisture content of the aggregate. This is reflected in the variation
,in slump as 'sholAmin Table A2. After the first batch was mixed} the amount of
water was adjusted'somewhat to correct the consistency if needed so that more
uniform concrete was obtained in later batches. The last batch was always
placed in the, critical stub area so that it was of 'knm,ffi strength for use in
analysis. Three or more 6 by'12-tn. test cylinders and a modulus of rupture
test beam were made from each batch . The properties of the ',c-oncrete are given in
Table A2.
One day after the test spec,i.mens were cast forms were
removed. Thespec,imen was covered with and surrounded by poly-
'ethylene plastic to hold in the moisture . This curing continued for one week,
following which the beam was ,left in the ,lab uncovered until tested. The
control cylinders and beams were cured in the same way and were tested on the
same day as the beam itself. Table A2 gives the age at'testing for the beam.
31. Details. of Test Apparatus and Instrumentation
The beams were tested in a test frame with the load. being applied.
by means of '50-ton hydraulic -rams. Figure 4 shows- a photograph of the test
frame with abeam in place. Since some beams were subjected to reversal of
'loadingJend supports carrying a reaction eJther up or down were
required. It was also desired that the beam be simply supported with freedom
to move horizontally'at the supports to accormnodate changes in length which
might oocur during the test.
-211-
The support system used consisted of an end yoke at each end with a
shaft of S .A.E'. 4340 heat treated s-teel extending 'out on each side of the beam
at mid-height. Holes were provided to accommodate 12) 16) and 20-in. deep
beams. Cam yoke rollers slipped onto the shafts to provide the desired s-imple
support condition. The end reaction supports were constructed in such a way
that a milled horizontal surface was provided for the cam yoke rollers to b.ear
against with loading in either direction. The top piec-e was not necessary for
'beams having only downward loading) but with this piece in place the end support
reaction consisted of as-lotted opening wi th the cam yoke roller 'in the middle
(Fig. 4). A thin piece of shim stock was 'placed temporarily between the top
of the roller -and. the top surface as the cap piece was bolted in ,place with the
splice plates, and upon.removal a slight clearance was 'maintained to allow
freedom of movement of the reaction horizontally.
The end reaction supports were securely 'bolted to a pair of steel
channels extending between the steel columns of the test frame. Column bases
wereoolted to tie-down sockets in the floor of the laboratory to provide
restraint against uplift of the columns. Thus the jack for upward loading
could bear directly on the floor of the laboratory. The jack for downward
loading acted against a cross beam carried on a pair of steel channels spanning
between columns of the test frame.
Most of the beams deflected several inches during the course of the
test to failure) but the 50-ton hydraulic rams used had a stroke length of
only 6 in. Jack extensions in various lengths were fabricated to allow inser-
tion of an extension with the jack retracted while the beam was unloaded) after
which the test was continued for -another -stroke length before repeating the
process. Since it was planned that each beam be unloaded at several points
during the test, this procedure worked out very well and required a minimum of
trouole to add the extension while other Teadingswere being taken.
-212-
The load was measured by means of dynamometers designed and built
for thisproject. Figure A5 shows across-sect.ion view of the load dynamometer
consisting of two cylinders 'marked ,IlAII andJ'B
II
threaded together -as
Parts nell andllDtl are not a part of the ,load measuring-system) but they trans-
,fer the applied load to cylinders ITA 11 and !lBTT by means of the threaded con ..
nections show-n. When a compressive load is 'applied to the dynamometer,the
1
inner 'cylinder It A" is in compress ion and the outer cylinder. "B" is in
tens-ion. Four Type A 7 -.4 SR-.4electric Btraingages (60 ohms Tes istance) were
mounted at positions 90 deg. apart on the outside surface of . "A.iI , . and four of
the same type gages were mounted at positions 90 deg. apart on the inside sur-
face of .ITB. IT These eight gages were then combined into a.fuIlWheatstone
bridge to provide maximum output, withtwb opposite tension gages or tw'O
opposite compression gages in each of the four arms of the The
cylinders "A" and llBlTwere constructed of heat-treated S .A. E. 4130 steel, and
the thicknesses were proportioned on the basis ofastress of 45ksi at the
60 kip capacity of the dynamometer. The sensitivity was 'approximatelY"lO Ibs
per micro-inch strain output from the bridge described above.
In most cases , .two of these load dynamometers were insert'ed .between
the jack and the beam" ,one being'connected to a portable strain indica tor'which
'wasused to measure the load at various stages during the test and the other
'connected to a plotter 'in .order to record the load continuously throughout the
test. For'some of the early tests" only one dynamometer was used" and a double
throw switch was installed to permit it to be switched from .thecontinuous
plotter '::t'ecordto the strain ,indicator 'when loading was stopped. This had the
disadvantage of neglecting any drop-off in load before .the dynamometer'Teading
was taken on the indicator" but the plotter -record could be used to estimate
the peak load obtained where this difference was significant.
-213-
The deflections were' measured by means of 0 .- OOl--in . dial indi ca t-ors
with 4-in. travel carried by a deflection bridge supported at the end bearings.
This gave a non--deflecting :reference fordeflecttons which were measured at
four points along the span. These dials bore against short pieces of angle
cemented to the fac-e of the beam at mid-height. Two of the readings were
taken at points in line with the faces of the stub and two midway between the
support and the face of the stub on either 'side. The dial indicators were
mounted on the deflection bridge in such a way that they could easily be res-et
when the 4-in. travel was about to be exceeded. When failure of the beams
having very large ultimate deflections was approaching, the deflection bridge
was -removed and deflection readings at midspan were read to the nearest 0.01 in.
with a scale.
An Autograph Model 3 x-yPlotterwas used to obtain a continuous
load-deflection record . The midspan deflection was measured by a slidewire-
displaeement device having:18--in. travel in either 'direction which was -con-
nected to the x-input of the plotter at all times during the test. A load-
dynamometer was connected to the y-inputas desc-r-ibed above. The support
bracket for -the displacement measuring device was mounted on the cross beam of
the test frame . For downward loading, this arrangement result-ed in larger
measured deflections on the plotter than the actual deflections, -since the
def-Iection;":of the test frame was included . Since thiserrorincreas-ed linearly
with load, it did not influence the shape of the load-deflection curve which
was the intended product of the continuous record. The more accurate dial
readings were used in analysis of the data, and the def1ection at any g-iven
point could thus be found without including the error from the test frame
deflection.
Strains in the tension and compression steel were measured at
several points by means of electrical strain gages mounted on the bars and
-214-
waterproofed before casting the concrete as described in Section 30. The leads
from these gages -which had been clearly marked before casting were connected to
a switch box. TheII'ype A7-4 SR-4 gages required an _ordinary Baldwin portable
strain indicator, while the .post-yield gages, TypePA-3 SR-.4, -reCluired a special

post-yield strain indicator ." The strain gage types and locations for the
different beamsare-given.in Table A3.
Strains' in the' concrete also l.reremeasuredwi th electrical resistance
strain gages. These gages were mounted at several points as indicated in
Table A3, the pattern 'during -the cours-e of the investigation .
However; one gage was always located one in. from the face of the stub on the
top surf'aceof the beam. The Baldwin portable strain indicator used for -the
steel strain readings was also used for these gages.
Strains were. measured on the sides of three beams ) .using a combination
of 2-in. and. 6-in. ,gage lengths .A2-in. Whittemore strain gage and a 6--in.
Berry strain gage were used for -these measurements until their ranges were
exceeded. Direet reading mechanical strain gages equipped vi th O. OOl--in. dial
indicators were used for higher 'Strains. Steel plugs w'i thgage holes drilled
in their -fac-eswerecemented to the surface of the concrete in the patterns
shown in Fig .A6 us ing Eastman 910 cement. The2--in. gage lengths were located
inthe'vicinity of the stub where the highest curvatures were expected, and the
overlapping 6-in. gage lengths extended some distance further out toward the
support.
32. Des cription of Test Procedure
The first step in preparing the beams for testing was to mount the
end yokes on the beam and place it in the test frame. Plaster of Paris was
* '
This point-yield indicator 'Was a _portable strain indicator which had been
modified to provide an external bridge -circui tfor measurement of -strains
larger thanthecapacityofthe usual range extender.
-215-
then used to seat the channel shaped pieces which formed the top and bottom
of these yokes once the beam was in place, care being taken to insure simul-
taneous 'contact between all four rollers and the end reaction supports.
After mounting the concrete strain gages, leads from thes'e gages as
well as the steel strain gages were hooked up to a switch box. The position
of . each gage was verified by balancing the gage on the strain indicator -and
observing the unbalance caused by shorting the gage with a shunt wire. An
initial set of readings was taken for 'all gages at zero load., and the plotter
'load and deflection calibrations were set for the scales desired atearlyloads.
The beam was loaded in increments with 50-ton hydraulic 'rams powered
by an electric pump . At every load interval, the load dynamometer reading and
deflection dial readings were taken. Steel and concrete strains were measured
at significant points during the test and photographs were taken to record the
appearance of the beam as cracking progressed. The load was removed and
reapplied 'at several points during the test to get a measure of the change in
stiffnes s of the member. After the early stages of the test , . the plotter
deflection scale was -reduced to one in. deflection equal to one in. on the
graph for the remainder of the test.
The concrete cylinders and beamsv.Tere test'ed on the same day the beam
was tested. Where tests extended over more than one day, these tests were made
on the day in which the beam was carried to first crushing -stage,which was the
firstday,in almost all cases. For 'some of the beams which were very ductile,
the testing extended over a' .period of two or three days. This was caused
partly by the 10 to 20 minutes required to take readings at each load level and
partly because of the half day schedule of testing time available in many cases.
Load was never allowed to remain.on the beam for periods longer than the time
required to take a complete set of readings.
-216-
TABLE Al
PROPERTIES OF REINFORCING BARS
(#8 Bars Except as Noted)
Beam
Tension Reinforcement
No. f E
Esh fult Eult
f E
Y
sy fract fracto
ksi
%, %
ksi
% % %
J-l 476
0.172
077 876 125
84.1
J-2 48.0 0.181' 096 873 845
J-3 48'3
0.170
075 885 125 871
J ~ __ ---- ____ ,--44. 9 0 .152 1.86 71.0 150 64.5 20.0
' ~ - . .., ... - - - - . - - - . - - - - . . . . . ~
"-_....... __. ..-:----
J-5 451
0.195 1
0
77 737
163 69.2 1905
3-6 46.2 0.178 1.84 ' 73.6 150 68.5 '
187
3-7 46.5 0.182
171 74.0 150
68.6,'
18.1
3-8 454
0.178 192 72.1 163, 670 ' 18.-8
3..,9 470 ' 0.175 1.66 7400 150 69.,3
195
3-10
451 0.195
179 730
15,0 68.0
185
J-11
469
0.188
157 75.4 150 705
18.8
3;..12
451
0.174 1.86 71.8 150 667
18.8
J-13 456
0.188 '1.61
733
150 6902 18.8
J-14 471
0.183
175
74.0 1,5.0 69.2
175
J':'17 ' 469 0.188 1.69 73.6 1 5 ' ~ 0
705
18.8
3-18 454
0.191 158 732 150 692 175
3-19 458
0.178 1.85 72.0 163 65.4 '18.8
3-20 458
0.185 176 71.8 163 65.4 183
J-21 476
0.186 1.80 72.1 1500
667
18.8
J-22 46.2 0.181 1.68
734 150 692 175
3-24* 485
0.180
190 778
16.0 68.1 24.1
*' , #4 bars.
-217-
TABLE Al (Continued)
PROPERTIES OF REINFORCING BARS
(#8 Bars Except as Noted)
Beam
Tension Reinforcement
No. f E
Esh fult Eult
f E
Y
sy fract fract
ksi
% %
ksi
% % %
J-2*
48,6 0.180
133
78.6 13.8
675 178
J-3
48.0 0.165
073 895 125
86.0
J-5* 489
0.170 1.22 82.6
125 78.4 138
J"'9
46.4 0.192
173 750
15.0 69.3 195
J-7 463
0.186
157 751
13.8
705
18.0
J-8 455
0.162
173 723 150
66.6 18.8
J-12* 497
0.180 1.45 823 125 794
14.0
J-13
46.0 0.173 198 71.8 150 667
18.8
J-14* 50.0 0.170 1.22 81.2
125 74.4
J-17
46.8 0.188
155 749 150 705 ' 18.8
J-18 471
0.191 1.68
747 150 705 175
J-20 46.5 0.180
174 737
150 692
18.8
J-22 46.4 0.161 1.83
723 150
66.6
175
J-24** 478
0.180
155
76.9 16.0
652
22.8
* #6 bars.
**
#4 bars.
-218-
TABLE A2
SUMMARY OF CONCRETE PROPERTIES
Beam Cylinder Strength, psi Slump, . in. Age
No.
Batch 1 Batch 2 Batch 3 Batch 4 Batch 5 1 2
3 4
5
Days
J-l 5700 5180 4930 05 25 25 50
J-ll 4370 4000 4110
05
2.0
15
8
J-2 4190 3870 4080
35 35 45 147
J-8 4280 4650 4680
65 45 45 27
J-17 4990 3900 15
2.0
39
J-3 4950 4680 4900
25 45 30 90
J-l0 3410 3890 3880
3590 50 25 50 50 15
J-14 5950 5100 4950 4500 1.0
30 35 50 29
J-13 4550 5100 5500 4800
45 30 30 35
22
J-19 3880 4030 3930 3900 05 15 15
1.0
17
J-20 4500 4460 4190 4380 1.0 1.0
15
1.0 18
J-4 5420 5550 5660 5520 4820 2.0 2.0 2.0
30
2.0
76
J-9 4720 4570 4390 4440 4190
25
4.0
50 55 55
19
J-5 5130 5030 4700 5010 5000 1.0 2.0
30 25 25 76
J-12 4700 4050 4450 5300 4550 1.0 400
30
1.0
30
41
J-6 5010 4990 4750 5440 5160
15 30 35 15
2.0 81
J-7
4220 4540 4470 4470 4470 4.0
35 30 30 35
19
J-21 4530 4500 4590 4350
1.0
15
1.0
15
20
J-22 4630 4660 4220 4420
15
2.0 400 2.0 22
J-18 4760 4410 1.0 1.0
34
J-24 4950 5000 60
-219-
TABLE A3
SUMMARY OF STRAIN MEASUREMENTS TAKEN
Beam
Noo
J-l
J-2
J-3
J-4
J-5
J-6
J-7
J-S
J-9
J-l0
J-ll
J-12
J-13
J-14
J-17
J-1S
J-19
J-20
J-21
J-22
J-24
*
Tension Steel
* Type Position
T
l
0,7,16
0,7,16
0,7,16
0,7,16
0,7,16
0,7,16
0,7,16,26
0,7,16,26
0,7,16,26
7,16)26
7)16)26
7,,16)26
7,16,26
7)16)26
7)16,,26
Compression Steel
* Type Position
~
None
7
7
None
7
7
0)7)16,,26
0)7,,16)26
None
None
None
0" 7"16,, 26
7,16,26
7)16)26
7,16,26
None
7
None
Distance from centerline of span.
** Distance from face of stub.
Concrete Gages
** Type Positiof.\
A-l
A-l
A-l
A-l
A-3
A-3
A-3
A-l
A-l
A-3
A-3
A-3
A-3
A-3
A ~ 3
A-3
A-3
A-3
A-3
A-3
A-l
1)6
1,4,7,10
1)4,7,10
1,4,7,10
1,4,7,10
1,4,,7,10
1)4,7,10
1
1
1
1
1
1
1
1)4,7,10
15, 20
1,4,,7,10
15, 20
1
1
1
1
1,4,7,10
15" 20
Gages on Side
of Beam
None
None
None
None
+
Type A- 3 Gages
+
Type A-3 Gages
+
TypeA-3 Gages
++
TypeA-l Gages
++
Type A -1 Gages
++
Type A-3 Gages
++
T y p e A ~ 3 Gages
++
TypeA-3 Gages
Mechanical Gages
Mechanical Gages
Mechanical Gages
None
None
None
None
None
Mechanical Gages
+ Gages 1-7/16 and 2 in. down from top at 1-1/2 and 6 in. out, respectively.
++
Gages 2 in. down from top at 1) 10" and 20 in, out.
"rl
(!l

til
Ul
(lJ
H

U)
90 ,----,
80
10
60
5.0
.40
30
20
10
o
o 1
I I I I I 1 I
I
_ .. J
i
,
--j
I

J


I
/
L

I
I
-.-
j
I
---f----
. I
1
1
J
I I
I
I I
I
1 I
2
3 4
5
6
7
8
9
10 11 12 13. 14 15 16 17
- Percent
FIG. Al TYPICAL STRESS STRAIN CURVE FOR REINFORCING BARS USED IN THE CURRENT TEST PROGRAM
I
I\)

I
-22l-
. (a) During Test with Extensometer MOunted on Specimen
(b) After Failure with Extensometer Removed
FIG .. P2 PHOTOGRAPHS OF SPECIMENS USED IN COMPRESSION
TEST OF REINFORCING BARS
..-I
to

I
I
to
t.O
(l)
H
+l
(f)
b
-'
oM
CfJ

I
CfJ
CfJ
CI)
H
+l
(f)
6
-.J
-222-
60
50
40
30
20
10
Compre
r\
,

...-
1------

IT-
I
I
L

I
'l//. v/..
I
No. B Bar 6
t1

-r///

0
r
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Strain--Percent
60
'tompre

,...


I
.... _-
" ----
50
40
30
20
10
.-

I
--
ens ion
1 "//-'/ v/.
I
No. p Bar
6"
I
-'--
I- /7.,,//' v/-

o 0 0.2 0.4 0.6 0.8 laO 1.2 1.4 1.6 1.8 "2.0
Strain--Percent
FIG .. A3 COMPARISON OF O-E CURVES IN TENSION AND COMPRESSION
i-1/2ft 1-1/2"
1-1/2"
(a) Without
Compression Steel
6"
111
H
1"
J
2tl
,.
... 223-
1-1/2"
d
d
1-1/2"
II III
d
It

.-r
Compress ion-
8"
1-1/2"
1-1/2
1t
-:=E-
1
/
2
'
d
D/2
(c) Beams With b = 6 in.
(d) Beams With U-Stirrups
Note: All closed stirrups welded at the top with
1-172 in.. lap of bar ..
2ft
1
FIG. A4 DETAILS OF STIRRUPS USED IN BEAMS OF THE CURRENT TEST PROGRAM
-224-
Slip Fi t
NOTE: PARTS A AND B ARE OF S.A.E. 4130 HEAT-'rREATED STEEL.
TYPE A7-4 SR-4STRAIN GAGES ARE MOUNTED ON OUTSIDE
OF A AND INSIDE OF B.
FIG. A5 CROSS SECTION OF LOAD DYNAMOMETER
D
-225-
(a) Pattern for Beam J-17
(b) Pattern for. Beams J-13 and J-14
FIG. A6 PHOTOGRAPHS SHOWING MECHANICAL STRAIN GAGE PLUG LOCATIONS
APPENDIX B--DESCRIPrION OF COMPUTER PROGRAMS
33. Purpose of Programs
The three computer programs which were written and used in this
study perfonned the computations given below utilizing the IRM 650 digital
computer.
Yield Stage Program computedkd, cp , E ,
Y c
E'
s'
E EM. and n
s' sy' y"
for a given reinforced concrete cross--section utilizing the slightly modified
straight line theory developed in Section 12.
(2) Crushing Stage Program computed kc
d
, CPe' ES' and Mc
fora given reinforced concrete c-ross--section utilizing the analysis
in Section 13.
Ultimate Stage Program com"ute:d f1 If" I' k dcp E' E M
'.t:' cu C ' b ' u' u -' s u -' S J U -'
andM (meas.)/M (comp.) fora given reinforced concrete cross-section utilizing
u u
theanalys
34. Input,Data'and Preset Quantities
T1,1einput .data contained on two cards for the Yield and Crushing
Stage Programs, a third card being required for the Ultimate Stage Program.
Cards No. land 2 contained the following quantities scaled as
Data Card No {all .. celreta Card No 2 (all: three .. programs)
First Word -Beam Number
-8
Second Word - b x 10 inl ..
Third Word -d x 10-
8
in .. '
Fourth Word - A x 10-
8

. s
Fifth Word - f x
10-
10
psi
y
Sixth Word f x
10-
10
psi
11
Seventh Word -E (f 130,000)
y y
Eighth Word - Esh
-226-
-82
First Word _AI x 10 in.
s
Second Word __ f' x 10-.
10
.
y PSl
Third 'Word - x 10 --10 psi
Fourth Word - E'
Y
Fifth Word
-6 ( )
N x 10' ,Sequence No.
Sixth Word
d' x in.
-10
Seventh Word - -f! x 10 psi
c
Eighth Word - -f If 1 X 10--5
cu c
-227-
-r5a-ta.- Ca:r:d No.3 CUI timate Stage Program only)
First Word bl! x 10-.8 in.
Second Wordd
ll
x 10-
8
in.
contained the following:
(Stirrup Width)
(Stirrup Depth)
Third Word
-A" x 10--8 in. 2 (Stirrup Steel Area)
s
Fourth Word -
Fifth W0rd n'/2 x 10-.8 in.
Sixth Word
,-.8. )
- -s x 10 in. (Stirrup Spacing
Seventh Word M (meas.) x
10-
5
ft.-kips
u
6. (meas.) x
10-
8
in.
u

For ,the Yield StagePr0gram the initial torial value for -kd .ispreset
to a value 'smaller ,than the expected final value. The program adds increments
to this initial value until thesum 0f.momerits 'of transformed area about the
neutral a.xis is zero. The additional strain increment of 0.0003 is -preset as
a constant for use in Eq. 8 of the analysis developed in Section 12.
Forthe Crushing Stage Program the limiting strain of o.004corre-
-sponding to crushing -is preset along with the initial trial value of'k d. The
c
reduction:factor, F, is preset as 0.85 for the analysis developed, in Section 13.
For the Ultimate Stage Program' the strain at the level of' the
compression steel, and the average ul t'imate concretestres s, f are
computed as 'functions of thebinding :ratio,Pb' as described inSect-ion 14.
The initial trial value of-kd is preset to a value smaller ,than the expected
u
kd. Empirical constants in connection with computationofp
b
, E' , and f"
u su cu
are preset to the values given in the equations developed in Section 14.
35 . Flow Diagrams
Yield Stage Program
Figure Bl gives the flow diagram for the Yield Stage Program. A
brief explanation of the various 'Steps is given below:
-228-
Step 1. After the input data are read from Cards 1 and 2 into
specified locations, set the initial trial value ofkd.
Step 2. Compute n and store.
Steps 3 through 6. Compute the moments of transformed areas about
the trial position of the neutral axis as indicated.
Step 7. Examine the algebraic 'sum of the moments of transformed
areas to see if it is close to being zero.
Step 8. If the unbalance lis large, inc-rease the initial value of'kd
by'Ool in. before repeating Steps 3 through 6 ~
Step 9. When the algebraic sum of moments of transformed areas comes
within a specified amount of being balanced, checkwhetheritis a negative
quantity.
Step 10. Add a finer increment of 0001 in. to kd until the value
correct to the nearest O. 01 in. is found as the value which makes the sign cif
the sum ofmaments of transformed areas become negative.
Stew 11 through 17 are self-explanatory 0 The various quantities
are stored in locations such that they will be printed out in the positions
g'i ven in Se cti on ~ 3 6 .
Crushing Stage Program
Figure B2 gives the flow diagram for the Crushing Stage Program.
A briefexp1anation of the various steps is given below =
'Step 1. Read the input data from Cards 1 and 2 into specified
locations" and modify addresses for storage of tensile and compressive forces
acting on the section.
Step 2.
. Step 3.
Step 4.
Set the initial trial value of 'k d.
c
Compute the average concrete stress at crushing -from Eq. 10 .
Set the li.rni tingstrain associated with crushing.
-.229-
St'ep 5.. Compute the curvature at crushing 'from Eq. 11 and store.
Step 6. .Examine to see if there is any compression 'st'eel, skipping
to Step. 11 if there is' none.
Steps 7 through 10 are self-explanatory. These steps compute the
compressive force in the steel if itis.present .
. St'ep 11. Compute the compressive force carried by -the concrete and
store.
Step 12. Compute the total compressive force and.store .
. Step 13. Using the preset reduction factor, F, compute the strain
in the tension steel and store.
Steps. '14 through:18 are self-explanatory. The force in the tension
steel is computed utiliz.ing:the generalized stress-strain 'curve of Fig. 27.
St:ep:: 19. Check to see iftensionand compression forces acting. on
the 'sect'ionare within 20 kips .of equilibrium.
Ste:g:; '20. Iftheunbalanceismorethah 20 kips, add 0.1 in .. to the
initial trial value 'for k 'dbefore 'Yepeating:steps 4 through 19 .
c
. Step 21-. If the unbalance is. less than 20 kips,check to see if (T-C)
is 'zero 0r negative.
Step 22. Increment kd until the neutral axis is. located .to the
c
nearest 0.01 in. by the change in.signof the algebraic 'sum of forces acting
on the section.
Step 23 Compute the crushing . Eq. 12 and store.
Step 24. Print the output in the positions listed in-Section 36,
ill timate Stage Program
Figure B3g'ives the flow diagram for theUltimate Stage Program. A
br-iefexplanation of the various steps isg-ivenbelow:
-230-
Step.l. Read data Cards 1) 2) and 3 and modify addresses for
stored Cluantities.
Step 2 is self -explanatory ..
SteEs :Lz,4,and 5 utilize equations developed in Section 14.
Empirical coefficients and constants are preset to values shown in these
expressions.
Step 6. Compute the force in the 'compression steel, neglecting
'strain hardening in thisanalys is J and store. (Note alternate path which may
be used if strain hardening 'is considered.)
Step 7. Set the initial trial value of k dsmallerthan the final
u
value. This is the distance from the level of compression 'Steel to the neutral
axis. For beams without compression steel this is referred to the point 2 in.
down from the original top fiber of the beam.
Steps 8 anc19 are self-explanatory.
Steps 10 through 15. Compute the strain in the tens ion st'eel and
the force in the tension steel utilizing the generalized stress-'strain curve
of Fig. 27. This quantity-is stored in a known location.
Step 16. Che'ck the equilibrium of tension and compression 'forces
acting on the section to see if they are within 20 kips of being balanced.
Step 17. Add 0.1 in. to previous trial k d before repeating Steps 8
u
through 16 if unbalance is more than 20 kips.
Step 18. When unbalance becomes less than 20 kips) see if the
section is balanced.
Step 12. Add increments to kd until it is located to the nearest
u
0.01 in. by the change in sign .ofthe unbalance of forces at the section.
SteI' 20. Compute M byEq. 20 and store.
u
Steps 21 and 22 are self-explanatory.
-231-
36. Output and 'Estimation of RunningTime
The output 'from the three programs was printed by the on-line print'er,
one line for 'eachheam analyzed. The arrangement of computed quantities and
scaling 'ineach case is givenbelow.
Yield Stage Program Output
First Word - Beam No.
W 10
-8
. Second 'ora - -kd x
Third Word- -cry
Fourth Word- 'E
c
,Fifth Word - 'E r
S
Sixth Word - 'E
sy
Seventh -Word ,...M x 10"".8
y
--.8
Eighth Word ,... n xlO
Ninth Word .- ,000
.y
Crushing -Stage Program Output
First Word - -Beam No.
Second Word,... (r jf:! }xlO-
5
eu e
Third Word-'E xlO--
6
c
Fourth Word- -kd x 10.,.8 in.
c
Fifth Word - 'CPc
Sixth Word - 'E'
S
Seventh Word -E
s
.... 8
Eighth Word ,... M x 10 in. -kips
c
Ultimate Stage Program Output
First Word - Beam No.
Second Word jf,g}x 10'-5
cu c
-232-'
Ultimate Stage Program .Out-put (continued)
Third Word - Pbx 1.0..,6
-8
Fourth Word - k d x 10 in.
u
Fifth Word - m
TU
Sixth Word - E'
su
Seventh Word - E
s
Eighth Word - M xlO -.8 ft. -kips
u
Ninth Word- Mu (meas . ) /Mu (comp .) x 10-.8
Estimation of Running Time
Eachof'the three computer programs used in this study'incorpo;rates
an initial est:i.rnate of' theposi tion of'the neutral axis. Increments of 0 .10 in.
are added to the initial trial valueatf'irst,then O.Ol-in. increments are
-added until the final position is located. This means that the running time
is closely related to the accuracy of this firsttrial value since this
determines the number of trials which must be made .
. Acormnoninitial trial value was used in the Yield StageProg;ram for
the beams of all three test programs. Some beams re'luired alonger running
time than others, but the average value was about five seconds per beam. Using
a comrnoninitial trial value for k d in the Crushing Stage Program,the average
c
time required was about 10 to 15 seconds. For the Ultimate Stage Program the
average running time l;ITas about 15 seconds per beam.
37. Availability
The three programs described above, including sample input and
output forms, coding sheets, operator's instructions, and the actual IBM 650
program listing ahd deck, has been placed in the Computer Program Library of
the Civil Engineering Department under File No. 268T. The yield and crushing
-233-
stage programs are separate decks. The ultimate stage program deck is made
up of the crushing stage program deck plus an additional deck which is
included separately.
, )
-
1..)
-234-
stut
I
b- input data
Sst;" kd (kd)
0
I
Oompute 11
:and (n-l)

I
16)lai .' lui ...
kd=kd+ 0 .. 01 J
I
3)Compute mom.oftranlformed
I
area of A! about n.utral axil
I
4)Compute moment of concrete
f
area about neutral axi.
I
total moment of tran.'l
area about n. & ..

'->Oomput. moment at transtormed
area of A. about neutral
I
(M .. '- M -
9) negative?
No
tens. compo
No
")
M 1lJ) negatiTel

il (M
t
-
ens. com ..
[ II) C ompu. t, E,1' trom Jqo 8 and store
I
I
llP)
store Beam No. in print location
1
B

Oompute
<Py and
store in print 100
I

1'+)comPUte

and &ltore in print
100. I
I
1'1)oompute O
and
"
store in print
locl
I
l't4JcomPUte M and store in prln t loc.
I
y
II
1 If)
Write
I
IIIIIIII
s tart ..
IIG" Bl now DIAGRAM lOB YIELD STAGI OOMPO'TJIR PROGRAM
I
0.10 I
I
Start -235-
,
I) Read input data and mod.i:f7
add.ressee tor quanti t
III .
Set for first trial
IJI
"5) Oompute f Jt
l
from Eq. gO ]
au c
and store in print location
4) Set c and store in print loc;'j.tion I
I
<Pc and store in
"II!!:S
___ III--_.II-c-____ -f_I_)-'C"-'o"-"m"",Crm"..."u"'-'t'-"!e'--'-'- 1 and 8 tore in ________ III'_I!lJ.. ____ _
l S
. "!ls (1- ') 0 T
II Y
NO
lIG.B2 trow DIAGRAM rOR CRUSHING STAGJ11 COr-rPUTaIR
... 236 ...
Slrt
I' )
R.ad inpu.t data and modify
I
addresses for Itorad quantities
I

store Beam NOa in print location
1
iii
I')
Compute P
b
and store in print 100 ..
I
III
1
4
)
Compute
t' It'
and Itore in print
100 .. 1
Oil. c
II!
I Compute

and store in print 100.
1
I
ill (' ..... ' ) :: 0 T
-,
I
au. sh
I
I<:;)co'lilpute
C
(yield) I
(atr.
hard .. ) r
..
i
___ ::J
I
I
17) c::.At
k d
:s
(k d)
I ...
u u 0

l I
kd
=
k d +
0.01 J
If
U u
1
05
) Compute
I
<Pc
and store in print loco
I
l-') Compute C and. store in print loc ..
I
c
T
['0) Compute
ts
and store in print loco
I
I
Ii)
( -
y)
:: 0 1 il
I
B
I
( .. J
iz (t -
sh)
,. 0
8

ri04-)
(Yield)]
lIe)
(Btl' .. hard.) I
T (alas t ic )
! Compute T
I C -0 IJ pu t e T
L.

I J
14
!Go)
(T om C
- 2C)
11)
0.10 l
I
is = O? k d
=
k d +
u
u
la)
('1 - C) il = 0 T
I .
rTo}-------------.. -.- -1------- -.-
L Compute M ana store in print loc.
-----------.-.j:!--.---.=r=-----.-----.---
E------------ ----.--------- 3
Compute M (measo)/M (compo) and store
u u
--------. I. ===-------- I
L----.-------!f---- . I
start
FIG-" B) FLOW DIAGRAM FOR ULTIMATE S1 WID COMPUr.::5IR PRGJRAM
APPENDIX C -LOAD-DEFLECTION C'JRVES
This appendix contains load versus midspan deflection curves for
beams of the current test program with the exception of three beams which
were subjected to reversed loading. Also included are load-deflection curves
for the beams from McCollister's test progrqm which were analyzed at all of
the four significant stages of behavior c.onsidered in this s t u d y ~ cracking,
yield, crushing, and ultimate. The average unloading-reloading curves are
shoT,lm and beam data are given in these figures .
For each beam, the plot of load versus deflection for the first two
inches is shown to an expanded scale in the top half of the figure. The
curve draT,lm to this scale shows the breaks at cracking and at yield for the
beam. Computed cracking and yield points are shown on the curves with X's.
Measured crushing is shown on these curves with a short cross-line and the
computed crushing point (using Method 1 for deflection) is shown with an X
for comparison.
The lower half of each figure shows the complete load-deflection
curve for thebeam,and the measured ultimate point is marked (V) on the
curve. The computed yield, crushing, and ultimate points are shown with X's
on these curves for comparison with the measured load-deflection response.
-237-
co
~
~
, ~
..,-(
~ ,
.3
.... 238 ...
BEAM DATA
b == 8.0 psi d
:: 10.0 in.
fl :: 4930 psi
pi
:: o %
c
f :: 47,600 psi p :: 1.98 'fo
y
50
C
10
I
\.
/"V
/"
~
//"
V
"
.-",/'
,
/ /-
//
//
.-
~ /
v
20
~
~ />' /./
V
///.,/
v/
.,./'
~ /
/;
J;-'
/' o
o 0.2 0.4 0.6 0.8, 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Midspan Deflection in Inches
50
40
30
('I
U I
'-J
L
20
10
~ I ..............
r J II ;
'"
I '
f
!
I
l L
! I
I
/. /
i
/
o
o
2'
4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG .. Cl LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-1
50
10
/
V ,/
V
v'/'
o
b
:::
f' :=
c
f :::
y
V/V
/
v/
8 .. 0 in.
... 239-
BEAM DATA
d = 10.0 in.
4080 psi pi:: 1 .. 10 'to
48,000 psi
p = 1.98 %
c
,IT 1\ /'
//
V/
//
V'
/ /'/
V
,/
./
V/.
/'
//
""
/""
o 0 .. 2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
40
U
....JL
C
V--
fJ"
-; X "
~ ~
fli
I
/
:; \
I
j
~ II
I / /
10
o
o 2 4 6 8 10 12 14 16 .18 20
Midspan Deflection in Inches
FIG. C2 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM J-2
-240 ...
BEAM DATA
b := 8.0 in ..
d
:::. 18.0 in ..
f':::. 4820 psi pI
:::.
-0 cfo
c
f :::. 44,900 psi p :::. ,1.10 %
y
50
c
.
)f
. -
/ r 7
20
V
/'
:/
/
1/
/
j
"
/1
I
/
/ 10
r/
!
1/
/ / " /

o 0 .. 2 O .. ~ 0.6 0.8 1.0 1.2 1 ~ 4 1.6 1.8 2 .. 0
Midspan Deflection in Inches
50
,
v
" C
~
~
~
~
-
I /\
I
I
!
I
I
/
I
I
i
I
-- /
10
o
o 2 4 6 8 10 12 14 16 18' 20
Midspan Deflection io locbes
,
FIG .. C3 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-4
I
-241-
BEAM DATA
b = 8.0 in. d = 18.0 in.
f' = 5000 psi pi = 0.61 %
c
f = 45
3
100 psi P = 1010 %
y
50
50
10
o
c
)c
I


/
/
rr /
//
/
/
/
/
/1
/
/
/
II
/
I
;//
/
--
/.
I II
I I til
o . 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.82.0
in-Inches
U
.. 'Jf

F""""
J
-,.;; (foR
CTUR!=')
......-

i
I
V
i
C
I
i
(11
i
I
i
I
I
I
j
r
I
I
I
/
i
!
I I
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. c4 LOAD vERSus MIDSPAB DEFLECTION FOR BEAM J-5
co

d

i1
.3
50
40
r7
J
VI
/
I;
,/
20
10
o
W/ /
/ /
-242 ...
BEAM DATA
b := 8.0 in ..
f' :: 5160 psi
c
f :: 46,200 psi
y
d := 18.0 in ..
pi :: 1.10 %
P == 1.10 %
- .
c
II
I
/
II
/
o 0.2 0.4 0.6 0.8 1.0' 1.2 1.4 1.6 1.8 2.0 2.2
Midspan in Inches
./
I
u
X .
I I .1 R"
50
I
.... T
"
V
V

i'
c

40
X ...... i

I
I /
\
I
30
I
I
I
/
I
I
I
I
20
I
/
10
II
f
0
I
0 2 4- 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C5. LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-6
-243-
BEAM DATA
b
== 8 .. 0 in. d
=
10.0 in.
f' = 4680 psi
pi
=
1.98 'fo
c
f
=
45,400 psi p
=
1,,98 %
y
50
1 c
~
. ,
.,.
/,,/
V/
/:
,/
,,/
v"/'
10
~ ./
//'
~
. ,./"
///
/
~ / /
//
o
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0'
Midspan Deflection in Inches
50
n
J
C
~ - ~ I
-;
--
rr
I
I
I
I
I
I
/
~ /
/'
I
II I
1/
x
20
10
o
o 2 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG.. c6 LOAD VERSUS MIDSPAN DEFIECTION FOR BEAM J-8
b :::
B.o in .. d ::: 1B .. o i n ~
I
f' ::
4190 psi P.'
::
CJ1,
C
f ::
47,000 psi
P
::
1 .. 10 '/0
y
50
c
)(
/ IT
~
/
V/
II
II
I
;
/
/
i
/
/
j,1
i
/
/
i
/
/
10
11 I /
1/
V 'j /
,I / I
/ / '/ /
o
o 0 .. 2 O .. ~ 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
50
40
~ 20
3
10
o
Midspan Deflection in Inches
. t
u
...
_J
-.
c
-
"-
(f'!
I
"
I I
I
I
I I
i
I
I
!
I
i
j'
!
I
I
I
I
/ i
i
!
I
i I I
o 2 4. 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG" C7 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-9
... 245-
BEAM DATA
b
:= 8 .. 0 in.
d :::: 14.0 in.
fi ::
3590 psi
pI
::::
o %
c
f
:::: 45,100 psi p :::: 1 .. 41 %
y
50
c
,
,
V
/
[7
/
20
V //
//
l : ~ / ' I II I I II /1V I I
50
~ 20
.3
10
o
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2 .. 0 2.2
Midspan Deflection in Inches
c
u
-
~
'\
I
I I
! I
o 2 6 8 10 14 16 18 20
Midspan Deflection in Incbes
FIG. c8 LOAD VERSUSMIDSPAN DEFLECTION FOR BEAM J-l.O
... 246-
BEAM DATA
b
=
8.0 in.
d
=
10.0 in.
f' :: 4110 pSi.
p' :: o %
c
f
=
46,900 psi
y
p
==
1.98 %.
50
c
'"
~
/.
//
/'
./
/
/"
.....
./
~ ,///
V
V
V'
~
// .
. / ~ > /
10
o
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches .
50
40
c
~
.Ir
r
A.
!
~
i
~
i
I I
.10
o
o 2 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C9 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-ll
... 247-
BEAM DATA
b
==
8.0 in. d
==
14.0 in.
f' :: 4800 psi
pI
==
1.41 %
c
f :: 45,600 psi p ::::I
1 .. 41 %
y
50
c
..
/;
?
1/
V
II
v
..
/
/
I
V ,/
/
/
/1
J
!
l
/ / J
l/
//
/
/
I
/
/ /
W/'I
J
I
v/
I
I
/1
/
/
j"
/
10
o
o 0.2 0.4 0.6 0.8 1.0. 1.2 1.4 1.6 1.8 2.0
50
40
il 20
.9
10
o
Midspan in Inches
u
y
"SHE:AR I
X
-
""
C
r-T

I
!\
"--
I
:1 /
I
I
I
I I Ii I I
n
I I
!
f
i
I
I
I
I!
I
II
I I I I
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG.C10 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-13
-248-
BEAM DATA
b
=
8.0 in.
d
=
14.0 in.
fl = 4500 psi
pI
= 0 .. 79 %
c
f =
47,100 psi
P =
1.41 ~
Y
50
c
-
/;
-:
,
V
i
II
I
10
V
,II
/
" /
.I
/
/
VI/
.I
/
/
/ o
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Incbes
50
u
"
r----
-;
\
.}(.,
C
~
~
\
.. I ...
40
) ~
i
I
,
I
I
i I
I
I
I
I
I
I
I i I
I !
I I
I
I
I
I
10
o
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG.ell LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM J -14
2.2
-249-
BEAM DATA
b
=
8.0 in. d
= 10.0 in ..
f' ::
3900 psi
pi
::
1.98 %
c
f
= 46,900 psi p :=
1.98 'to
y
50
40
CD
~ 30
~
d
"1"1
c
]
20
10
V
//'
//
/
v/
v//"
/ /
./"
//
./
,.. ./
V
7
~
. ~ / ' .
//
/'
{/':/
/
/'
l,./ .
//
//
o
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
40
U
01 r
i1 20
3
C
!
II
l
II"
II
rrf
I
II
II
I
'\
/
I /
/ /
;
II
/
!
l 7 7 I I I
10
0
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Incbes
FIG. C12 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J -17
... 250 ...
BEAM DATA
b
==
6 .. 0 in .. d ::: 10.0 in.
ft =:
4410 psi
p' == 2.63 %
c
f ::: 45,400 psi p
==
2 .. 63 %
y
50
J I I I I
I i
I I
c
20
~ '
.. ~ .
V
/ ./'
V
~ /
/
V'
10
o
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
u
~
X
C
~
r--
/ / ?
~
-
n ~
I / / /
~ I
I
il
1/
/ / /
10
o
o 2 6 8 10 12 14 ,16 18 20
Midspan Deflection in Inches
FIG It C13 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM J -18
-251-
BEAM DATA
b
= 8.0 in. d
==
14.0 in.
fl s
3900 psi
pi ::
o %
c
f ==
y
45,800 psi p
==
1 .. 41 %
50
C
I
.
"
/
-
/ /
VI
II
V
/ /
20
10
o V V II
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
,
C
~
t\
20
10
o
/
-
o 2 6 8 10 12 14 16 18 20
Midspan Deflect,ion in Inches
FIG. C14 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J -19
b
=
f' -=
c
f
==
y
- 2 5 2 ~
BEAM DATA
8.0 in. d
:= 14.0 in.
4380 -psi pi == 1.41 %
45,800 psi _
P
== 1.41 %
50
C
//
'I'
I
VI
I- /
o 0.2 0.4 0.6 0.8 1.0 1.2 l.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
u
X
.......
L..! "S?II APr."
~
~
rT
C
~
~
~
I
~
I
'/
I
20
10
o
I I
o :2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG" C15 LOAD VERSUS MIDSPAB DEFLECTIOI FOR BEAM J -20
"
50
40
n
J
~ /
Ii
V
f /
10
o
... 253-
BEAM DATA
b = 8.0 in.
f'! - 4350 psi
-c -
f :::: 47,600 psi
y
C
)(
I
I
V
/
I
. d == 18 .. 0 in.
p' :::: 0
p :::: 1.10%
,
-
,0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
50
40
10
o
o
Midspan Deflection in Inches
T
,..
,X
v
.....sIIfIIIII -.
;-Y
~ ,
I
2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG.. c16 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J -21
~
~
~
s:3
-rl
rd
~
o
~
to
P4
-rl
~
~
-rl.
rd
cd
0
H
50
40
30
20
10
III
J
'/ I
//
I
o '/
-254-
BEAM DATA
b = 8.0 in ..
f' = 4420 psi
c
f = 46,200 psi
y
. !
I
I
I
r
I
!
d = 18.0 in.
pI = 1.10%
p = 1 .. 10 'to
Ie
rx
I
/
!
/
/
V
/
i
j
I
o 0.2 0.4 0.6 0.8 l.O 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
u
x
50
40
30
-'
5I.ff:AR"
~ ~
IJiIiiI"'"
~
I
I
!
......
......
~ I
I
X .....
~
I'
!
20
10
I
I
t
I I
I
I
I
i
I
!
I
I
o
1
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C17 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM J-22
-255 ...
BEAM DATA
b
=
6 in. d
=
10.58 in.
f' ::
3900 psi
p'
= 0 .. 63 %
c
f = 41,800 p ~ i
p
= 1039 %
y
50
40
X
~
v
10
V
o
V
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
50
40
-il 20
3
10
o
Midspan Deflection in Inches
U
J ,
~ I
'J
~
""
~
"
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. c18 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM T-1
-256 ...
BEAM DATA
b
=
6 .. 0 in. d
=
10 .. 37 in.
f' :::
3860 psi
pi
:::
1 .. 93 %
c
f
=
45,400 psi p
==
3,,21 %
y
50
40
C
A-
X;
"
I
10
/
o
V
o 0.2 0.4 0.6 0.8 1.0 1 ~ 2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
u
J
./
V--
><
\
:;---
V
\
10
o
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C19 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM T-2
-257-
. BEAM DATA
b
= 6 in. d
= 10.20 in.
tl ==
4270 psi
p'
== 2 .. 58 % /
c
f
= 44,700 psi p
= 5 .. 10 %
y
c
50
v
1,/
i
!
~
V
/
20
10
o
V
o 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
u
~
50
c V l ~ \ I
jL L I
)
}I"
\
20
10
o
o 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C20 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM T-3
... 258-
BEAM DATA
b
::
6.0 in. d
==
10.65 in.
fl :=
2230 psi p'
== 0 .. 63 rf;
c
f
==
47,100 psi
P == 0.97 rf;
y
50
40
C
'X,
10
v_
I
o
VV
o 0 .. 2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Midspan Deflection in Inches
50
40
u
,....
-\.
v
-------
~
~
10
o
o 2 4 6 8 10 12 14 16 18 20
Midspan Def1ection in Inches
FIG. C2.1 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM T-4
-259-
. BEAM DATA
b
==
6.0 in. d
=
10.51 in.
tl ::
2020 psi
pI
==
0 .. 98%
c
f :: 48,400 psi p
==
1.90 %
y
50
40
w
~ 30
~
s:=
c
. "1""'1
]
20
10
~
Iv
.,
/
V
o
V
o 0.2 0 .. 4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
50
~ . 20
3
10
o
Midspan Deflection in Inches
p
~ ~
~
o 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C22 LOAD VERSUS MIDSPAN DEFLECTION FOR BEAM T-5
... 260-
BEAM DATA
b :: 6 .. 0 in ..
f' == 1910 psi
c
d :: 10 .. 37 in ..
pi:: 1.93 %
f == 56,500 psi p:: 3,,21' %
y



10 /
of
o 0.2 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
_ ......_ ......_/11 .... _ .............._. __ ......._._ ... _._ ...._ ....__ ...._ ..... _ .._ ..................__ ......__ .... ... .._ .. ... ... .... _. ___ .....__ - ...---------...- ... - .....- ....-.........- ...........-......._.-.-........ --
50
u
40

I
" \
m
t1
30
:2
oM

20
3
10
0
0 2 4 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. C23 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM T-6
-261-
BEAM DATA
b
==
6.0 in. d := 10.58 in. .t
f' := 4-540 psi
pI
:=
1.39 %
c
f := 40,900 psi p :=
1.39 %
y
50
40
CD
~ 30
~
g
..-i
~
20
I.
V!
.. I
10
V
0 V
0 0 .. 2 0 .. 4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2 .. 0
5 0 - - - - - - - - - - ~ - - ~ - - ~ - - - - ~ - - ~ - - - - ~ - - ~ - - - - ~ - - ~
20 ( ~
l:rll
FrFtcture
I II - I I I I I
024 6 8 10 12 14 16 18 20
Midspan Deflection in Inches
FIG. c24 LOAD VERSUS MIDsPAN DEFLECTION FOR BEAM T-7